Book Review: A Borderlands View on Latinos, Latin Americans, and Decolonization: Rethinking Mental Health

Aborderlands view on Latinos, Latin Americans, and decolonization: rethinkingmental health by Pilar Hernandez-Wolfe central thesis focuses on what the author terms “intersectionality”. Intersectionality can be thought of as the interplay of race, ethnicity, gender, sexual orientation, migration, class, language, and macroeconomic and social-ecological influences on our lives. The author, a licensed marriage and family therapist, helps readers climb a virtual mountain of historic underpinnings of social injustice, and arriving at the peak, we are able to understand how this foundation of intersections form the basis of an identity and frequently serve as the basis of individual or familial mental health pathology.

Rather than blame the individual, the author holds superpowers such as the United States, and colonializers, such as Spain, as the responsible parties for much of the suffering in the Americas. Her definition of colonialization is that colonization happens any time an outside culture invades an existent culture, resulting in a loss of knowledge of and appreciation for the existent culture, and giving rise to a new, third culture that lacks understanding of the original culture and ways of being. From this perspective, colonization is a macro-traumatic act, forcing entire societies into submission and giving rise to a new social order where the colonized people internalize globalization and perpetuate colonization against their own people.

People who have internalized inequality and unhelpful bystanders come to perpetuate inequality in their own societies, and learn only upon immigration to the United States, that even if they are considered “White” or  of the ruling class in their country of origin, in the U.S., Latinos are frequently assimilated into the class that is “ruled over”.

Drawing on feminist principals of sharing of economic and social power, the author stipulates that the pathway towards healing lies first in developing critical conscience. Only when people begin to understand the exterior forces that led them to points of crisis in their own lives, can they begin to see inequities in society that harm both the individual as well as anyone they are exercising “power over”. The author explores “Just Therapy” and “Transformative Family Therapy” as theoretical models for treating mental health issues in minority clients. However, I would suggest that beyond looking only at minority clients, the treatment models suggested could be useful for clients of all backgrounds. While White middle class heterosexual clients might not go to marriage therapy looking for someone to “blame” for their problems, the real answer lies in examining personal history and the exposures (biological, social, educational) that have laid the basis of our automatic cognitions. One of the reasons I hypothesize that Cognitive Behavioral Therapy has been indicated for so many conditions is that it is goal focused and practical and its techniques are replicable.

Just Therapy and Transformative Family Therapy may be more difficult to replicate, as one of the central approaches involves males and females breaking into gender congruent groups to discuss the issues that brought them to therapy and develop a conscious awareness of the societal and economic forces that prevail in their lives and may have been the germination bed of their inter-familial issue. Numerous therapists are used and the therapeutic practice itself meets regularly to discuss each therapists’ praxis of the racial, ethnic, gender, sexual-orientation, and class biases that may impacts his or her own ability to deliver Just/Transformative Therapy.

This approach, with its critical analysis of race, ethnicity, gender, sexual identity, class identity, and migration history may not prove helpful for people from highly individualistic societies, or those that are drawn to solution focused therapeutic approaches. However, especially for people who come from societies with a collective orientation, this approach may be more helpful than CBT.

One of the essential take aways from this book is of the need for therapists to examine their own privileges. Rather than identify “White” privilege as the culpable issue, the author takes on the shades of beige and the ways that Latinos in the U.S. may use their “minority” status as a way to rule over Latinos of a “lesser” class. Arguing that everyone has some privilege, the author urges readers to think about all the different ways that their particular race, ethnicity, gender, sexual-orientation, and education has shaped their perception of the world and the ways that these particular facets of identity elevate or lower how they are perceived in society, and how that perception affects the power distance they experience between themselves and their clients.

While the author weaves together theoretical and treatment approaches from a variety of the social sciences, there are portions of the book that would have benefited from the loving hand of a copy editor and additional critical feedback from peers. If the purpose of the chapter that contains information on Just Therapy and Transformative Family Therapy was to provide instruction on how to implement these therapeutic modalities, then the chapter could really benefit from addition of more material that at minimum could provide the standard “Plan-Do-Study-Act” approach used in dissemination sciences. It seemed to me that the foundational planning and strategizing mental health practices would need to undertake in order to implement this approach wasn’t sufficiently covered in a way that would facilitate planning to undertake this type of practice transformation.

This book is informative for graduate level students seeking to deepen their understanding of the historic issues inherent in the mental health treatment of Latinos in the United States. The study guide at the end provides questions for critical consideration and could serve as a study guide for a seminar on Latino mental health. However, this book does not offer typical prevalence or incidence information for students seeking to understand basic information on Latino mental health in the United States. Additional reading would be required to understand the scope and burden of specific mental health problems of Latinos in the U.S.

Checking the Box: A Graduate School Essay on Race and Ethnicity

The question of whether or not I am White is easy for me to answer. All standardized forms with a question about my race and ethnicity have been checked [X] White, [X] Non-Hispanic since I was a child.

Generally, there is never a box that asks:
“Did you grow up in a rural area?”
“Did you ever have health insurance as a child?”
Or, “Have you ever lived in poverty?”. Since I am White, whatever adverse circumstances I might have faced growing up are now irrelevant to how I get treated at the Department of Motor Vehicles or how long I wait at the doctor’s office.

White privilege is real, and how I perceive it is mostly based on the absence of White privilege that my husband faces. Need a title to a car we bought? It is my job to go to the DMV because, “The people at the DMV never suspect White people are trying to get fake titles.”
Need a doctor’s appointment? It is my job to call, make the appointment, and explain our insurance to both the doctor’s office and my husband. “Nobody understand me on the phone, and I don’t understand our health insurance,” my husband states.

My husband never got asked about whether to be White or not growing up in Bolivia. Seventy percent of Bolivians are Indigenous, and three tribes, the Quechua, Aymara, and Guarani account for the majority of the population. The government collects no official statistics on “race” since the underlying assumption is that everyone born in Bolivia is Bolivian and the laws that govern land ownership are associated with tribal law and tribal territory. Hence, if a person resides in a particular geographic area, association with a tribe is established by being born in that particular place. Instead of race, all national identity cards come with a specific drill down on where people are born. Based on where he was born, my husband is presumed to be Guarani or a mix of Guarani and Quechua. As an accountant, he didn’t need farm or grazing lands, it didn’t really matter. The question of my husband’s race was first posed to us as a choice by the U.S. State Department.

“Ponga blanco, ponga Latino” he said, (Put down I’m White, put down I’m Latino).
Later on, when we were living in the United States, and our Census form came, I asked him how he would like to record his race.
“Hay una cajita para mestizo?” (Is there a little box for mixed?)
“There is”, I replied.
“Selecione esa.” (Pick that.)
“Quieres que ponga que sos latino?” (Do you want me to put down that you are Latino?), I asked.
“Si, esta bien.”

The question of what race he felt he was came up again in pregnancy, when the time came to make a choice about genetic screening. We didn’t need screened for whatever genetic problems Ashkenazi Jews have. Did we want to be screened for cystic fibrosis, which is common in peoples of European ancestry? Or did we want to be screened for sickle cell which was more common in people of color? We decided that whatever the genetics between us, the error rate on the tests was too high to end a pregnancy over the results, so we preferred not to have the tests.

On June 9, 2014, our lives changed forever when Carlos was born. Bolivians, in addition to wanting to know where you are born, want to know who the parents are. Bolivia extends citizenship to children born abroad and also mandates that all children shall be given a paternal last name and a maternal last name. So, when it came time to name our son, I put down my name, my husband’s names, and our son’s names.

Following a parallel procedure to fill out the application in the state of Maryland, a huge box appeared on the Maryland Department of Vital Statistics Web site: “WARNING! None of the names on this birth certificate match! Do you wish to continue?”

Yes, I did wish to continue. I had written the names, checked “White” for race, and checked “Latino or Hispanic” for ethnicity.

In 2014, it seemed like this was a perfectly normal and reasonable thing to do, as I lived in my DC metro area bubble, where one in four people are foreign born and I was imagining an America of inclusion and possibility. What did it matter to have ridiculously long sets of names and to check the Latino box?

More difficult to answer than race and ethnicity is the question, “What language is mainly spoken at home?” This is a challenge because we are swimming in a language soup of words that represent our different realities when we are together and when we are apart. When my husband first came to the U.S., he didn’t speak any English, and the first sentence I taught him in English was “I am a legal permanent resident.” It is really hard to learn another language and to practice it all day long. It is hard to lose your independence and your profession and your place in society. And so, in the beginning, because he didn’t understand, and because he was exhausted, we spoke Spanish at home 100% of the time.

As time has gone by, he has learned to speak English, and his English is fluent enough to work in a retail environment. Our son can speak both English and Spanish, but we mainly speak Spanish at home because it feels comfortable, normal and simple. But when it comes time to check the box on forms, I hesitate. Will people think my son is “behind” because of the language we speak at home? Will they assume a level of ability or inability based on his name and the boxes I have checked?

One box we are happy to be able to check is the “U.S. Citizen” box. I’m not sure why, but the recent Executive Action banning immigrants from Muslim majority countries enrages me. Maybe it’s the mountains of visa paperwork I have filled out in my lifetime. Maybe it’s the fact that immigrants, through the processing fees they pay, fund United States Citizen and Immigration Services operations. It feels so unjust, so unfair, that after years of filling out forms and going to interviews and getting your fingerprints taken and getting tax records from sponsors and getting your Tuberculosis titer level checked (I’ll just state a second time that these folks paid for this stuff!!), that someone in power could just arbitrarily decide you don’t qualify because people where you had the luck of being born check the “Muslim” box on a question of religious faith.


The rage I feel about the Muslim ban is something new for me. Maybe it was the images of the drowned toddler being carried by the Turkish soldier. Maybe it’s the fact that no matter which boxes of race, ethnicity, or religion we feel we best fit into, we all want things like good health and safety for our children. The America I thought I knew was better than this. The America I thought I was a part of, was a country where everyone could find a place and no person was illegal. It makes me think that trips to the DMV notwithstanding, I’ve been blind to the injustice that continues to exist in America. The question I ask myself now is, “What do we to change the system,” so that we can move toward a society where any box checked is ok, and all checked boxes are equally valued.

Graduate Level Health Economics Econometric Solutions to Woolridge Computer Problems C6.4 C6.10 9.3 9.4 C9.1 C9.8

. use "E:\data\gpa2-1.dta"

. regress sat hsize hsizesq

      Source |       SS       df       MS              Number of obs =    4137

-------------+------------------------------           F(  2,  4134) =   15.93

       Model |  614822.097     2  307411.048           Prob > F      =  0.0000

    Residual |  79759024.2  4134  19293.4263           R-squared     =  0.0076

-------------+------------------------------           Adj R-squared =  0.0072

       Total |  80373846.3  4136  19432.7481           Root MSE      =   138.9

------------------------------------------------------------------------------

         sat |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

       hsize |   19.81446   3.990666     4.97   0.000     11.99061    27.63831

     hsizesq |  -2.130606    .549004    -3.88   0.000    -3.206949   -1.054263

       _cons |   997.9805   6.203448   160.88   0.000     985.8184    1010.143

------------------------------------------------------------------------------

. * Yes the quadratic term is statistically significant

. *The optimal high school size

. gen logsat = ln(sat)

. regress logsat hsize hsizesq

      Source |       SS       df       MS              Number of obs =    4137

-------------+------------------------------           F(  2,  4134) =   16.19

       Model |  .614405203     2  .307202602           Prob > F      =  0.0000

    Residual |  78.4287724  4134  .018971643           R-squared     =  0.0078

-------------+------------------------------           Adj R-squared =  0.0073

       Total |  79.0431776  4136   .01911102           Root MSE      =  .13774

------------------------------------------------------------------------------

      logsat |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

       hsize |   .0196029   .0039572     4.95   0.000     .0118445    .0273612

     hsizesq |  -.0020872   .0005444    -3.83   0.000    -.0031546   -.0010199

       _cons |   6.896029   .0061515  1121.03   0.000     6.883969    6.908089

------------------------------------------------------------------------------

. *using the logsat as the DV the optimal high school size is 469

. clear

. use "E:\data\bwght2-1.dta"

. regress lbwght lbwght npvis npvissq

      Source |       SS       df       MS              Number of obs =    1764

-------------+------------------------------           F(  3,  1760) =       .

       Model |  74.2054098     3  24.7351366           Prob > F      =       .

    Residual |           0  1760           0           R-squared     =  1.0000

-------------+------------------------------           Adj R-squared =  1.0000

       Total |  74.2054098  1763   .04209042           Root MSE      =       0

------------------------------------------------------------------------------

      lbwght |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

      lbwght |          1          .        .       .            .           .

       npvis |  -7.94e-19          .        .       .            .           .

     npvissq |  -2.30e-20          .        .       .            .           .

       _cons |  -1.78e-15          .        .       .            .           .

------------------------------------------------------------------------------

. regress lbwght npvis npvissq

      Source |       SS       df       MS              Number of obs =    1764

-------------+------------------------------           F(  2,  1761) =   19.12

       Model |   1.5771321     2  .788566048           Prob > F      =  0.0000

    Residual |  72.6282777  1761  .041242634           R-squared     =  0.0213

-------------+------------------------------           Adj R-squared =  0.0201

       Total |  74.2054098  1763   .04209042           Root MSE      =  .20308

------------------------------------------------------------------------------

      lbwght |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

       npvis |   .0189167   .0036806     5.14   0.000     .0116979    .0261355

     npvissq |  -.0004288     .00012    -3.57   0.000    -.0006641   -.0001934

       _cons |   7.957883   .0273125   291.36   0.000     7.904314    8.011451

------------------------------------------------------------------------------

. *Yes the quadratic term is statistically significant

. tab nvpis

variable nvpis not found

r(111);

. tab npvis

      total |

  number of |

   prenatal |

     visits |      Freq.     Percent        Cum.

------------+-----------------------------------

          0 |          5        0.28        0.28

          1 |          2        0.11        0.40

          2 |          5        0.28        0.68

          3 |         12        0.68        1.36

          4 |          6        0.34        1.70

          5 |         27        1.53        3.23

          6 |         59        3.34        6.58

          7 |         58        3.29        9.86

          8 |        117        6.63       16.50

          9 |         96        5.44       21.94

         10 |        199       11.28       33.22

         11 |        115        6.52       39.74

         12 |        618       35.03       74.77

         13 |         72        4.08       78.85

         14 |         97        5.50       84.35

         15 |        143        8.11       92.46

         16 |         41        2.32       94.78

         17 |         12        0.68       95.46

         18 |         15        0.85       96.32

         19 |          4        0.23       96.54

         20 |         35        1.98       98.53

         21 |          5        0.28       98.81

         22 |          2        0.11       98.92

         23 |          1        0.06       98.98

         24 |          2        0.11       99.09

         25 |          3        0.17       99.26

         26 |          1        0.06       99.32

         30 |          7        0.40       99.72

         33 |          1        0.06       99.77

         35 |          1        0.06       99.83

         36 |          1        0.06       99.89

         40 |          2        0.11      100.00

------------+-----------------------------------

      Total |      1,764      100.00

. gen visits = .

(1832 missing values generated)

. replace visits = 1 if npvis >= 22

(89 real changes made)

. tab visits

     visits |      Freq.     Percent        Cum.

------------+-----------------------------------

          1 |         89      100.00      100.00

------------+-----------------------------------

      Total |         89      100.00

. tab npvis if npvis >21

      total |

  number of |

   prenatal |

     visits |      Freq.     Percent        Cum.

------------+-----------------------------------

         22 |          2        9.52        9.52

         23 |          1        4.76       14.29

         24 |          2        9.52       23.81

         25 |          3       14.29       38.10

         26 |          1        4.76       42.86

         30 |          7       33.33       76.19

         33 |          1        4.76       80.95

         35 |          1        4.76       85.71

         36 |          1        4.76       90.48

         40 |          2        9.52      100.00

------------+-----------------------------------

      Total |         21      100.00

. tab npvis visits

     total |

 number of |

  prenatal |   visits

    visits |         1 |     Total

-----------+-----------+----------

        22 |         2 |         2

        23 |         1 |         1

        24 |         2 |         2

        25 |         3 |         3

        26 |         1 |         1

        30 |         7 |         7

        33 |         1 |         1

        35 |         1 |         1

        36 |         1 |         1

        40 |         2 |         2

-----------+-----------+----------

     Total |        21 |        21

. *Yes it does make sense in that if there are thatmany visits it could be a problemat

> ic pregnancy with the likelihood of lower birth outcomes

. regress lbwght npvis npvissq mage magesq

      Source |       SS       df       MS              Number of obs =    1764

-------------+------------------------------           F(  4,  1759) =   11.56

       Model |  1.90136387     4  .475340968           Prob > F      =  0.0000

    Residual |  72.3040459  1759    .0411052           R-squared     =  0.0256

-------------+------------------------------           Adj R-squared =  0.0234

       Total |  74.2054098  1763   .04209042           Root MSE      =  .20274

------------------------------------------------------------------------------

      lbwght |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

       npvis |   .0180374   .0037086     4.86   0.000     .0107636    .0253112

     npvissq |  -.0004079   .0001204    -3.39   0.001    -.0006441   -.0001717

        mage |    .025392   .0092542     2.74   0.006     .0072417    .0435423

      magesq |  -.0004119   .0001548    -2.66   0.008    -.0007154   -.0001083

       _cons |   7.583713   .1370568    55.33   0.000     7.314901    7.852524

------------------------------------------------------------------------------

. *The optimal age is 31

. gen age = .

(1832 missing values generated)

. replace age = 1 if mage > 31

(605 real changes made)

. replace age = 0 if mage <= 31

(1227 real changes made)

. tab age

        age |      Freq.     Percent        Cum.

------------+-----------------------------------

          0 |      1,227       66.98       66.98

          1 |        605       33.02      100.00

------------+-----------------------------------

      Total |      1,832      100.00

. regress bwght npvis npvissq mage magesq

      Source |       SS       df       MS              Number of obs =    1764

-------------+------------------------------           F(  4,  1759) =    8.59

       Model |  11376019.1     4  2844004.78           Prob > F      =  0.0000

    Residual |   582383777  1759  331087.992           R-squared     =  0.0192

-------------+------------------------------           Adj R-squared =  0.0169

       Total |   593759796  1763  336789.448           Root MSE      =   575.4

------------------------------------------------------------------------------

       bwght |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

       npvis |   37.47386   10.52531     3.56   0.000     16.83042     58.1173

     npvissq |  -.7862864   .3417884    -2.30   0.022     -1.45664   -.1159322

        mage |   81.60554   26.26395     3.11   0.002     30.09369    133.1174

      magesq |  -1.327179   .4392829    -3.02   0.003     -2.18875   -.4656073

       _cons |   1860.381    388.977     4.78   0.000     1097.475    2623.286

------------------------------------------------------------------------------

9.3

       Let math10 denote the percentage of students at a Michigan high school receiving a passing score on a standardized math test (see also Ex 4.2).  We are interested in estimating the effect of per student spending on math performance.  A simple model is

i)       You need to be “poor” to be in the federally funded student lunch program.

ii)      Because lnchprg is negatively correlated with log(expend).  It is significant.

iii)    Yes

iv)      Math10 would decrease by 3.24% if lunch program increase by 10%.

v)       It’s a good predictor variable because R-squared is increasing.

9.4

(i)

Tvhours= tvhours*+e0

(ii)

It’s not likely to hold in the example, because if the tvhours=0 then they would be reported as zero

So the error depends on the actual tvhours

C9.1

(i)

. use "C:\Users\sphl\Desktop\CEOSAL1-1.DTA"

. generate rosneg = 0

. replace rosneg = 1 if (ros<1)

(23 real changes made)

. regress lsalary lsales roe rosneg

      Source |       SS       df       MS              Number of obs =     209

-------------+------------------------------           F(  3,   205) =   28.81

       Model |  19.7902019     3  6.59673397           Prob > F      =  0.0000

    Residual |  46.9319613   205  .228936397           R-squared     =  0.2966

-------------+------------------------------           Adj R-squared =  0.2863

       Total |  66.7221632   208  .320779631           Root MSE      =  .47847

------------------------------------------------------------------------------

     lsalary |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

      lsales |   .2883868   .0336172     8.58   0.000      .222107    .3546665

         roe |   .0166571   .0039681     4.20   0.000     .0088336    .0244806

      rosneg |   -.225675    .109338    -2.06   0.040    -.4412462   -.0101038

       _cons |   4.297602   .2932526    14.65   0.000     3.719425     4.87578

------------------------------------------------------------------------------

. ovtest

Ramsey RESET test using powers of the fitted values of lsalary

       Ho:  model has no omitted variables

                 F(3, 202) =      1.07

                  Prob > F =      0.3614

** We accept the null-hypothesis

C9.8

   1.  New update available; type -update all-

. use "C:\Users\sphl\Desktop\twoyear.dta"

. codebook stotal

--------------------------------------------------------------------------------------

stotal                                                   total standardized test score

--------------------------------------------------------------------------------------

                  type:  numeric (float)

                 range:  [-3.3247969,2.2353656]       units:  1.000e-09

         unique values:  227                      missing .:  0/6763

                  mean:   .047483

              std. dev:   .853544

           percentiles:        10%       25%       50%       75%       90%

                          -1.10531  -.327343         0   .610791   1.13706

. reg stotal jc

      Source |       SS       df       MS              Number of obs =    6763

-------------+------------------------------           F(  1,  6761) =    1.03

       Model |  .752020193     1  .752020193           Prob > F      =  0.3097

    Residual |   4925.6191  6761  .728534108           R-squared     =  0.0002

-------------+------------------------------           Adj R-squared =  0.0000

       Total |  4926.37112  6762  .728537581           Root MSE      =  .85354

------------------------------------------------------------------------------

      stotal |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

          jc |    .013658   .0134431     1.02   0.310    -.0126946    .0400107

       _cons |   .0428543   .0113348     3.78   0.000     .0206344    .0650741

------------------------------------------------------------------------------

. reg stotal univ

      Source |       SS       df       MS              Number of obs =    6763

-------------+------------------------------           F(  1,  6761) = 1574.72

       Model |  930.653211     1  930.653211           Prob > F      =  0.0000

    Residual |  3995.71791  6761  .590995106           R-squared     =  0.1889

-------------+------------------------------           Adj R-squared =  0.1888

       Total |  4926.37112  6762  .728537581           Root MSE      =  .76876

------------------------------------------------------------------------------

      stotal |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

        univ |   .1615085     .00407    39.68   0.000       .15353    .1694869

       _cons |  -.2636267   .0122004   -21.61   0.000    -.2875434     -.23971

------------------------------------------------------------------------------

. reg lwage jc univ exper stotal

      Source |       SS       df       MS              Number of obs =    6763

-------------+------------------------------           F(  4,  6758) =  500.23

       Model |  367.406832     4  91.8517079           Prob > F      =  0.0000

    Residual |  1240.88926  6758  .183617825           R-squared     =  0.2284

-------------+------------------------------           Adj R-squared =  0.2280

       Total |  1608.29609  6762  .237843255           Root MSE      =  .42851

------------------------------------------------------------------------------

       lwage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

          jc |   .0630514   .0068214     9.24   0.000     .0496792    .0764235

        univ |   .0686405   .0025651    26.76   0.000     .0636121     .073669

       exper |   .0048773   .0001571    31.04   0.000     .0045692    .0051854

      stotal |   .0493766   .0068096     7.25   0.000     .0360277    .0627255

       _cons |   1.495271   .0212176    70.47   0.000     1.453678    1.536864

------------------------------------------------------------------------------

. test jc + univ <0

+ not found

r(111);

. test jc+ univ<0

+ not found

r(111);

. test jc < univ

< not found

r(111);

. test jc + univ < 0

+ not found

r(111);

. reg lwage jc totcoll exper stotal

      Source |       SS       df       MS              Number of obs =    6763

-------------+------------------------------           F(  4,  6758) =  500.23

       Model |  367.406831     4  91.8517079           Prob > F      =  0.0000

    Residual |  1240.88926  6758  .183617825           R-squared     =  0.2284

-------------+------------------------------           Adj R-squared =  0.2280

       Total |  1608.29609  6762  .237843255           Root MSE      =  .42851

------------------------------------------------------------------------------

       lwage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

          jc |  -.0055892   .0069385    -0.81   0.421    -.0191909    .0080126

     totcoll |   .0686405   .0025651    26.76   0.000     .0636121     .073669

       exper |   .0048773   .0001571    31.04   0.000     .0045692    .0051854

      stotal |   .0493766   .0068096     7.25   0.000     .0360277    .0627255

       _cons |   1.495271   .0212176    70.47   0.000     1.453678    1.536864

------------------------------------------------------------------------------

. gen stotalsq = stotal * stotal

. reg lwage jc univ exper stotal stotalsq

      Source |       SS       df       MS              Number of obs =    6763

-------------+------------------------------           F(  5,  6757) =  400.17

       Model |  367.436832     5  73.4873664           Prob > F      =  0.0000

    Residual |  1240.85926  6757  .183640559           R-squared     =  0.2285

-------------+------------------------------           Adj R-squared =  0.2279

       Total |  1608.29609  6762  .237843255           Root MSE      =  .42853

------------------------------------------------------------------------------

       lwage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

          jc |   .0632092    .006833     9.25   0.000     .0498143    .0766041

        univ |   .0685131   .0025846    26.51   0.000     .0634466    .0735797

       exper |   .0048776   .0001572    31.04   0.000     .0045695    .0051857

      stotal |   .0501562   .0070778     7.09   0.000     .0362814     .064031

    stotalsq |   .0019191   .0047481     0.40   0.686    -.0073886    .0112268

       _cons |    1.49399   .0214545    69.64   0.000     1.451932    1.536047

------------------------------------------------------------------------------

. gen jcstotal = jc*stotal

. gen univtotal = univ*stotal

. reg lwage jc univ exper stotal jctotal univtotal

variable jctotal not found

r(111);

. reg lwage jc univ exper stotal jcstotal univtotal

      Source |       SS       df       MS              Number of obs =    6763

-------------+------------------------------           F(  6,  6756) =  334.24

       Model |  368.126196     6  61.3543661           Prob > F      =  0.0000

    Residual |   1240.1699  6756  .183565704           R-squared     =  0.2289

-------------+------------------------------           Adj R-squared =  0.2282

       Total |  1608.29609  6762  .237843255           Root MSE      =  .42845

------------------------------------------------------------------------------

       lwage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

          jc |    .063636   .0068983     9.22   0.000     .0501132    .0771588

        univ |   .0689706   .0027439    25.14   0.000     .0635917    .0743494

       exper |   .0048856   .0001572    31.08   0.000     .0045775    .0051938

      stotal |   .0582108   .0087001     6.69   0.000     .0411559    .0752657

    jcstotal |  -.0168818    .009275    -1.82   0.069    -.0350638    .0013001

   univtotal |  -.0026652   .0029894    -0.89   0.373    -.0085254     .003195

       _cons |   1.495924   .0212432    70.42   0.000      1.45428    1.537567

------------------------------------------------------------------------------

. test jcstotal univtotal

 ( 1)  jcstotal = 0

 ( 2)  univtotal = 0

       F(  2,  6756) =    1.96

            Prob > F =    0.1410

. * not jointly significant

. reg lwage jc univ exper stotal

      Source |       SS       df       MS              Number of obs =    6763

-------------+------------------------------           F(  4,  6758) =  500.23

       Model |  367.406832     4  91.8517079           Prob > F      =  0.0000

    Residual |  1240.88926  6758  .183617825           R-squared     =  0.2284

-------------+------------------------------           Adj R-squared =  0.2280

       Total |  1608.29609  6762  .237843255           Root MSE      =  .42851

------------------------------------------------------------------------------

       lwage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]

-------------+----------------------------------------------------------------

          jc |   .0630514   .0068214     9.24   0.000     .0496792    .0764235

        univ |   .0686405   .0025651    26.76   0.000     .0636121     .073669

       exper |   .0048773   .0001571    31.04   0.000     .0045692    .0051854

      stotal |   .0493766   .0068096     7.25   0.000     .0360277    .0627255

       _cons |   1.495271   .0212176    70.47   0.000     1.453678    1.536864

------------------------------------------------------------------------------

. * the original regression is the preferred one

.