Changes

Jump to navigation Jump to search
no edit summary
{{Project|Has project output=Content|Has sponsor=McNair ProjectsCenter|Project TitleHas title=Women in Entrepreneurship (Issue Brief)|Topic Area=Social Factors in Entrepreneurship|OwnerHas owner=Carlin Cherry|Start TermHas start date=Spring 2016|End TermHas keywords=Women, Entrepreneurship|StatusHas project status=Tabled|Deliverable}}=Timeline and plan for summer=*note: I am using "hopethisworks8.txt" to make my graphs*4-5 artifacts, 4-5 specifications, last artifact a regression table (tables, charts)**something about data itself. pie chart (how many companies do we have, how many do we have ceo information for, how many do we have founder) and then relatedly many can we classify as men or women***this will influence all remaining graphs - use this info to guide the rest of your graphs***classified drs using their first name matched to common first name list, that's why we were able to classify some drs as women, some as men, and some drs but gender unknown**graph with time on x axis, % on y axis, women founders, women ceos, women management positions (management position being anything that is vp and above) over time***will write up these results-ie most of the women in "womens management positions" **conditional on having (ceo identified) how many women, how many men, drs, dr men, dr women, dr unknown. **bar graph with women in various industries, y axis will have percentage***should also make note of total percentage of VC dollars in each industry**regression table***variable names, with n and r^2, industry/year fixed effects Y/N***each one of these will correspond to a regression***variables: IPO, acquisition, exit all correspond to 0,1***rounds, dollars invested =Spring 2017===Current plan to complete project==#Lit Review (which will be compiled here [[Women in Entrepreneurship Lit Review]]). Keep track of the things that have already been done in other research reports, so that we can do something different/original or so we can prove their data wrong. Done 3/28.#Brainstorm ways to fix Dr. problem in data (assigning gender to those who are doctors), recode them. reran new tables with new coded doctors. done 4/4.#Think of new variables to add depending on the guidance from the lit review (might include adding different variables than solely womenceo/womenfounder but will depend on litreview). ##add control variable for the people in data who have no gender#I think from there I can do analysis and start to write report! ==For analysis== Note that:*finaldatasetcode.sql (in 181/Women) has been updated*hopethisworks5.txt (in both 181/Women and E/Women) has been created**Use this as your dataset now**It contains IPO, IPOyear, IPOAmount, MA, MAyear, MAAmount, and Exit**IPO, MA, and Exit are 0/1 variables. Year is an int, and Amount is a real.**Many MAs will not have amounts. Build graphs of:*Over time (last ten yrs)**percent of co's with woman ceo (binary)**percent of co's with woman founder (binary)**percent of co's with woman clevel (binary)**percent of co's with women in top management (vp and above)**average fraction of women founders**average fraction of women clevel**average fraction of women in top management (vp and above)*Some of the above by industry*Some of the above by state Regresions:*rounds womanvar w/controls*inv womanvar w/controls*exit womanvar w/controls*ipo womanvar w/controls*ma womanvar w/controls*ipoamount womanvar w/controls*maamount womanvar w/controls ==Call with SDC Platinum to determine information about the data==#How does Thompson get their data?##Why is coverage better for some firms than others?##Is the data self-reported by the companies?#How does Thompson upload executive data?##How often – after last round of funding?##After it’s uploaded, is the data updated continuously?#Better coverage for companies that get more rounds of financing/IPO/M&A?  #VC data sourced from government filings, public press releases, and quarterly surveys of private equity firms. If company does not participate in survey, then SDC does not have the data, which is why coverage for some firms is better than others. #Deals team is the one that uploads all the data. As soon as SDC gets an update, or has a source that is updated, they automatically upload that to system as well. It typically takes 24-48 hours for new info to be reflected in database. #Yes, there is better coverage for companies that have IPO/M&A/get more rounds of financing. ==TO DO after 2/28==#Check cleancos and verify that it is actually clean (yes)#Make Dr. a control variable (yes)#Sum all c-level, etc.#aggregate to company level We need to build:*IPO 1/0*M&A 1/0*Number of rounds*Total invested*Is the CEO a woman 1/0*Are any of the founders women 1/0*Is CEO doctor 1/0*Number of founders*Number of founders that are doctors*Number of women founders*State - coded*Industry -code them!*Year of First investment -extracted Ed is going to add IPO and Acq to rc1andcp3 Here's the code to add zeros to cleanpeople: DROP TABLE cleanpeople2; CREATE TABLE cleanpeople2 AS SELECT prefix, firstname, lastname, jobtitle, fulltitle, cname, CASE WHEN genval IS NOT NULL THEN genval ELSE 0::int END AS women, CASE WHEN doctor IS NOT NULL THEN doctor ELSE 0::int END AS doctor, CASE WHEN charman IS NOT NULL THEN charman ELSE 0::int END AS charman, CASE WHEN ceo IS NOT NULL THEN ceo ELSE 0::int END AS ceo, CASE WHEN cfo IS NOT NULL THEN cfo ELSE 0::int END AS cfo, CASE WHEN coo IS NOT NULL THEN coo ELSE 0::int END AS coo, CASE WHEN cio IS NOT NULL THEN cio ELSE 0::int END AS cio, CASE WHEN cto IS NOT NULL THEN cto ELSE 0::int END AS cto, CASE WHEN otherclvl IS NOT NULL THEN otherclvl ELSE 0::int END AS otherclvl, CASE WHEN boardmember IS NOT NULL THEN boardmember ELSE 0::int END AS boardmember, CASE WHEN president IS NOT NULL THEN president ELSE 0::int END AS president, CASE WHEN vp IS NOT NULL THEN vp ELSE 0::int END AS vp, CASE WHEN founder IS NOT NULL THEN founder ELSE 0::int END AS founder, CASE WHEN director IS NOT NULL THEN director ELSE 0::int END AS director FROM cleanpeople WHERE jobtitle IS NOT NULL; ==TO DO up to 2/28== #Load##company##people##title lookup##state lookup#Refine people##gender 0/1##Join title lookup##Dr.'s?#Aggregate to company level (left join)##Agg. People##Join state lookup#Export! Done! Here is the code for this part of project: DROP TABLE titlelookup;CREATE TABLE titlelookup( fulltitle varchar(150), charman int, ceo int, cfo int, coo int, cio int, cto int, otherclvl int, boardmember int, president int, vp int, founder int, director int);\COPY titlelookup FROM 'Important Titles in Women2017 dataset.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV--628  DROP TABLE statelookup;CREATE TABLE statelookup( statename varchar(100), uniquecode int);\COPY statelookup FROM 'uniquestateval.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV--50  DROP TABLE people;CREATE TABLE people( prefix varchar(5), firstname varchar(50), lastname varchar(50), jobtitle varchar(150), cname varchar(150) );\COPY people FROM 'pull5-normal.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV--186181  DROP TABLE cleancos;CREATE TABLE cleancos ASSELECT pull3info.*FROM pull3infoJOIN Cleancosbase ON pull3info.companyname=cleancosbase.companyname AND minfirstdate=firstinvestdate;--43534  DROP TABLE genvalpeople;CREATE TABLE genvalpeople ASSELECT *,CASE WHEN prefix='Ms' THEN 1::intWHEN prefix='Mrs' THEN 1::intWHEN prefix='Mr' THEN 0::intELSE Null::int END AS genval FROM people;--186181 DROP TABLE cleanpeople;CREATE TABLE cleanpeople AS SELECT genvalpeople.*, titlelookup.*FROM genvalpeople LEFT JOIN titlelookup ON genvalpeople.jobtitle=Issue Brieffulltitle;--186181 DROP TABLE uniquevalstate;CREATE TABLE uniquevalstate ASSELECT cleancos.*, statelookup.*|AudienceFROM cleancos LEFT JOIN statelookup ON cleancos.companystate=General Publicstatelookup.statename;--43534  DROP TABLE dataset;CREATE TABLE dataset ASSELECT cleanpeople.*, Policymakersuniquevalstate.*FROM cleanpeople LEFT JOIN uniquevalstate ON cleanpeople.cname=uniquevalstate.companyname;--186181
|Keywords=Women, Entrepreneurship\COPY people TO 'finaldataset.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV|Primary Billing=AccMcNair01}}--186181
[[Women in Entrepreneurship Lit Review]]==Paper==https://docs.google.com/document/d/1RHC0Namt9VpmUxINmRvHw4U34dDmgYPcUMbwhhthcCA/edit
=Spring 2017===To-doTimeline==
*Prior
**review/update lit review
**state fix-effects
*Policy
**Trump's proposed policy changes that will impact vc-backed companies, entrepreneurship/innovation in general, women in the workplace: research can be found at [[Trump and Women in Entrepreneurship]]***crowd-funding impact
==Timeline==
*1/24 - finish outline, reread tables/data collection process, update lit review
*1/25 - SDC pull, rerun through tables w new added variables (i.e. industry, etc)*1/27 - read over new data and refine into something presentable, find appropriate parameters for issue brief, etc. maybe start making charts and graphs if time. @ this point check in w ed
*1/30 - finish rerun through tables w new added variables (i.e. industry, etc)*1/31 - read over new data and refine into something presentable, find appropriate parameters for issue brief, etc. maybe start making charts and graphs and chartsif time. @ this point check in w ed*2/1/31 - finish making graphs and charts, research trump policy and write it up*2/1 3 - start finalizing paper, putting all the pieces of the puzzle together, any additions or time to fix things that went off schedule*2/3 - , last edits, submit paper to anne for review
==SDC Search: Women2017=====pull 1===
*Database Selection
**VentureXpert
*Search Items:
**Date that received company received first investment: 01/01/1980 to 01/01/2017
**Moneytree Deals: Select All Moneytree Dealsventure-related deals
**Company Nation: U.S.
*Custom Report
**Company Founding Date
**Company IPO Date
**Date Received Last Investment
**Date Received First Investment
**Company Name
**Company State
**Total amt invested in company
**No. Financing Rounds Company Received
**Company Stage Level 1 at Round Date
**Company Stage Level 2 at Round Date
**Company Stage Level 3 at Round Date
**Executive's First Name
**Executive's Last Name
**Executive's Job Title
**Executive's Prefix
**VE Primary Industry Sub-Group 3
**VE Primary Industry Minor Group
**VE Primary Industry Sub-Group 1
**Company IPO
**Company Website
 
 
*Saved in 182 in mcnair/projects/women2017 as "pull1"
 
===pull 3===
*Database Selection
**VentureXpert
**Portfolio Companies
 
*Search Items:
**Date that received company received first investment: 01/01/1980 to 01/01/2017
**Venture related deals - select all venture related deals
**Company Nation: U.S.
 
*custom report
*company founding date
*company ipo date
*date company received last investment
*date company received first investment
*company staet
*company name
*company industry class
*company industry major group
*company industry subgroup 1
*company industry subgroup 3
*Company industry minor group
*company cusip
*total known amt invested in company
*no rouds company received investment
*company ipo status
*company website
 
*saved in 182 in mcnair/projects/women2017 as "pull3"
 
===pull 4===
*Database Selection
**VentureXpert
**Portfolio Companies
 
*Search Items:
**Date that received company received first investment: 01/01/1980 to 01/01/2017
**Venture-related deals
**Company Nation: U.S.
 
*custom report
*company name
*executive's name prefix
*executive's first name
*executive's last name
*executive's job title
 
*saved in 182 in mcnair/projects/women2017 as "pull4"
 
===pull 5===
*Database Selection
**VentureXpert
**Portfolio Companies
 
*Search Items:
**Date that received company received first investment: 01/01/1980 to 01/01/2017
**Venture-related deals
**Company Nation: U.S.
 
*custom report
*executive's name prefix
*executive's first name
*executive's last name
*executive's job title
*company name
*saved in 182 in mcnair/projects/women2017 as "pull5"
 
===pull 6===
*Database Selection
**VentureXpert
**Portfolio Companies
 
*Search Items:
**Date that received company received first investment: 01/01/1980 to 01/01/2017
**venture-related deals
**Company Nation: U.S.
 
*Custom Report
*executive's name prefix
*executive's first name
*executive's last name
*executive's job title
*company name
*company city
*company state
 
===pull 7===
*Database Selection
**VentureXpert
**Portfolio Companies
 
*Search Items:
**Date that received company received first investment: 01/01/1980 to 01/01/2017
**venture-related deals
**Company Nation: U.S.
 
*Custom Report
*executive's name prefix
*executive's first name
*executive's last name
*executive's job title
*company name
*Company ve primary industry major group
*Company ve primary industry minor group
*Company ve primary industry sub group 1
*Company ve primary industry sub group 2
*Company ve primary industry sub group 3
 
 
==What I have done up to 2/7/17 (in english, not sql)==
 
I have done many SDC pulls with many different variables. The most important tables are nowomenstate, nowomencity, nowomenminorind, and nowomensub2. nowomenstate shows the number of women in "power positions" in each state as a percentage of all power position workers. nowomencity does the same thing at a city level. nowomenminorind shows the number of women in power positions grouped by their company's minor industry group. nowomensub2 does the same except grouped by the company's industry sub group 2.
 
==Tables built up to 02/07/17==
 
SDC pull #1
company founding date
company ipo date
date company received last investment
date company received first investment
company staet
company name
company industry class
company industry major group
company industry subgroup 1
company industry subgroup 3
Company industry minor group
company cusip
ttoal known amt invested in company
no rouds company received investment
company ipo status
company website
 
SDC pull #2
company name
executive's name prefix
executive's first name
executive's last name
executive's job title
 
 
 
logging on:
researcher@128.42.44.181
password: 9million
 
cd /bulk
cd \Women2017
psql Women2017
 
DROP TABLE pull3info;
CREATE TABLE pull3info(
/*taken from pull3.txt*/
ipodate date,
lastinvestdate date,
firstinvestdate date,
companyfoundingdate date,
ipostatus varchar(10),
companystate varchar(50),
companyname varchar(200),
industryclass varchar(200),
majorgroup varchar(200),
minorgroup varchar(200),
subgroup1 varchar(200),
subgroup3 varchar(200),
cusip varchar(6),
totalinvested float,
numrounds int,
website varchar(200)
);
 
\COPY pull3info FROM 'pull3-normal.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV
 
--COPY 43606
 
DROP TABLE cleancosbase;
CREATE TABLE cleancosbase AS
SELECT companyname, min(firstinvestdate) as minfirstdate FROM pull3info WHERE companyname !='Undisclosed Company' OR companyname !='New York Digital Health LLC' GROUP BY companyname;
--43534
 
DROP TABLE cleancos;
CREATE TABLE cleancos AS
SELECT pull3info.*
FROM pull3info
JOIN Cleancosbase ON pull3info.companyname=cleancosbase.companyname AND minfirstdate=firstinvestdate;
--43534
 
 
DROP TABLE pull4info;
CREATE TABLE pull4info(
/*taken from pull4.txt*/
execfirstname varchar(50),
jobtitle varchar(100),
execlastname varchar(50),
execprefix varchar(5),
companyname varchar(200)
);
 
\COPY pull4info FROM 'pull4-normal.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV
 
/*the job title column of the exec is for some reason combined with the first name
even after normalizing the data. I will repull the data from SDC to see if it was
simply an error in the pull or something. If that doesn't work, will go to ed.
*/
 
SDC pull #5
executive's name prefix
executive's first name
executive's last name
executive's job title
company name
 
DROP TABLE pull5info;
CREATE TABLE pull5info(
/*taken from pull5.txt*/
execprefix varchar(5),
execfirstname varchar(50),
execlastname varchar(50),
jobtitle varchar(100),
companyname varchar(200)
);
 
\COPY pull5info FROM 'pull5-normal.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV
 
--COPY 186181
 
DROP TABLE distinctnocompanies;
CREATE TABLE distinctcompanies AS
SELECT DISTINCT companyname FROM pull3info;
/*created this per ed's advice to see how many companies are distinct from each other
so we dont get duplicates in the dataset - only really creating a table to get the
specific number*/
 
--COPY 43536
/*this means there are 70 duplicates in the data. could mean that some companies in diff
states have the same name, or it could be a for real duplicate. can go back and check
on this later but rn seems like a relatively nominal amt.*/
 
 
DROP TABLE totaltable;
CREATE TABLE totaltable AS
SELECT * FROM pull5info LEFT JOIN pull3info ON pull5info.CompanyName = pull3info.companyname;
/*represents the table combining pull3info and pull5info, aka all of the info about
the companies and the executives' info matched on company name*/
 
/*WHERE IM AT:
 
START GROUPING BY WOMEN IN pull3info
ed says to rename pull3info but idc like i know what it is
 
*/
 
DROP TABLE commonnames;
CREATE TABLE commonnames(
/*common male and female names with corresponding 0 or 1*/
name varchar(100),
mf int
);
 
DROP TABLE commonnamesclean;
CREATE TABLE commonnamesclean AS
SELECT name, sum(mf) AS mf FROM commonnames GROUP BY name HAVING count(mf) = 1;
 
\COPY commonnames FROM 'realcommonnames.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV
--4654
 
DROP TABLE impoppl;
CREATE TABLE impoppl AS
/*selecting the people with important job titles*/
SELECT * FROM pull5info WHERE jobtitle IN ('Board Member', 'CEO', 'Founder', 'CFO',
'COO', 'Chairman', 'Chief Executive Officer', 'Chief Financial Officer', 'President',
'Treasurer', 'Vice President');
--67733
 
DROP TABLE binarymanwoman;
CREATE TABLE binarymanwoman AS
/*creates table with a column populated with 1 if prefix is Ms or Mrs and 0 if mr*/
SELECT*,
CASE WHEN execprefix='Ms' THEN 1::int
WHEN execprefix='Mrs' THEN 1::int
WHEN execprefix='Mr' THEN 0::int
ELSE Null::int END AS binaryval FROM pull5info;
--186181
 
DROP TABLE binarycommonmatch;
CREATE TABLE binarycommonmatch AS
/*matching common names to list of execs who have dr prefix/are gender unidentified*/
SELECT execprefix, execfirstname, execlastname, jobtitle, companyname,
CASE WHEN binaryval IS NULL AND mf IS NOT NULL THEN mf
WHEN binaryval IS NOT NULL THEN binaryval
ELSE NULL::int END AS binaryval
FROM binarymanwoman LEFT JOIN commonnamesclean ON execfirstname = name
WHERE execfirstname IS NOT NULL;
--171307
 
DROP TABLE doctors;
CREATE TABLE doctors AS
/*represents the leaders from companies who haven't been assigned a gender*/
SELECT * FROM binarycommonmatch WHERE binaryval IS NULL;
--7016
 
 
DROP TABLE women;
CREATE TABLE women AS
SELECT * FROM binarycommonmatch WHERE binaryval=1;
--18101
/*this means that 18101/171307 of the people in "important" positions are women
which is equal to 0.105. Could be +/- 0.04 bc of the doctors/unidentified genders
in the data*/
 
/*next steps: redo SDC pull, group by
 
state
city
company
exec prefix, first name, last name, job title
also check in w ed about the no of doctos and whether getting within a 4% margin is
ok*/
 
SDC pull #6
 
company state
company city
company name
exec first name
exec last name
exec prefix
exec job title
 
DROP TABLE execbylocation;
CREATE TABLE execbylocation(
state varchar(75),
city varchar(75),
companyname varchar(100),
execjob varchar(100),
execprefix varchar(5),
firstname varchar(50),
lastname varchar(50)
);
 
\COPY execbylocation FROM 'pull6-normal.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV
--186242
 
 
DROP TABLE impopplbylocation;
CREATE TABLE impopplbylocation AS
/*selecting the people with important job titles*/
SELECT * FROM execbylocation WHERE execjob IN ('Board Member', 'CEO', 'Founder', 'CFO',
'COO', 'Chairman', 'Chief Executive Officer', 'Chief Financial Officer', 'President',
'Treasurer', 'Vice President');
--67750
 
DROP TABLE impoppllocname;
CREATE TABLE impoppllocname AS
SELECT * FROM impopplbylocation WHERE execprefix IN ('Ms', 'Mrs');
--5646
 
/*here's what I want to do: I have a table with all the (identifiable) women, and a
separate table with both men and women. I want to group by state in both, then
divide the no of females in state (impoppllocname)/total in state (impopplbylocation)
where state=state (i.e. females in california in impoppllocname divided by
total no of ppl from california in impopplbylocation) to get the percentage of females
in each state with "important" positions. however, sql is a turd and won't let me group
by state for some odd reason in either of these tables. is there a more efficient
way to do this than with two tables? perhaps. But idk! help!*/
 
DROP TABLE binpplbylocation;
CREATE TABLE binpplbylocation AS
/*creates table with a column populated with 1 if prefix is Ms or Mrs and 0 if mr*/
SELECT*,
CASE WHEN execprefix='Ms' THEN 1::float
WHEN execprefix='Mrs' THEN 1::float
WHEN execprefix='Mr' THEN 0::float
ELSE Null::int END AS binaryval FROM impopplbylocation;
--67750
 
DROP TABLE nowomenstate;
CREATE TABLE nowomenstate AS
SELECT binpplbylocation.state, sum(binaryval)/count(binaryval) AS percentage
FROM binpplbylocation GROUP BY binpplbylocation.state
ORDER BY percentage ASC;
/*this gives the no of women in each state/all the ppl in a state to yield the percentage
of women in each state who are females in power*/
--54 (includes Guam, D.C., etc.)
/*Wyoming the lowest, at 0, and Hawaii the highest, at 16.9%, Texas at 8.1%*/
 
DROP TABLE nowomencity;
CREATE TABLE nowomencity AS
SELECT binpplbylocation.city, sum(binaryval)/count(binaryval) AS percentage
FROM binpplbylocation GROUP BY binpplbylocation.city
ORDER BY percentage DESC;
--2049
 
SELECT * FROM nowomencity WHERE city='San Antonio';
--Houston has 7.8% of women in leadership positions in all of its companies
--Austin has 9.1%
--Dallas has 9.9%
--San Antonio has 7.4%
 
/*next do sdc pull to group by industry minor group 3, industry major group*/
 
SDC pull #7
 
Custom Report
executive's name prefix
executive's first name
executive's last name
executive's job title
company name
Company major group
company minor group
company sub group 1,2,3
 
perl NormalizeFixedWidth.pl -file="pull7.txt"
 
DROP TABLE execbyindustry;
CREATE TABLE execbyindustry(
execprefix varchar(5),
firstname varchar(50),
lastname varchar(50),
jobtitle varchar(50),
companyname varchar(100),
majorgroup varchar(100),
minorgroup varchar(100),
subgroup1 varchar(100),
subgroup2 varchar(100),
subgroup3 varchar(100)
);
 
\COPY execbyindustry FROM 'pull7-normal.txt' WITH DELIMITER AS E'\t' HEADER NULL AS '' CSV
--186272
 
DROP TABLE impopplbylocation2;
CREATE TABLE impopplbylocation2 AS
/*selecting the people with important job titles*/
SELECT * FROM execbyindustry WHERE jobtitle IN ('Board Member', 'CEO', 'Founder', 'CFO',
'COO', 'Chairman', 'Chief Executive Officer', 'Chief Financial Officer', 'President',
'Treasurer', 'Vice President');
--67764
 
DROP TABLE binpplbyindustry;
CREATE TABLE binpplbyindustry AS
/*creates table with a column populated with 1 if prefix is Ms or Mrs and 0 if mr*/
SELECT*,
CASE WHEN execprefix='Ms' THEN 1::float
WHEN execprefix='Mrs' THEN 1::float
WHEN execprefix='Mr' THEN 0::float
ELSE Null::int END AS binaryval FROM impopplbylocation2;
--67764
 
DROP TABLE nowomenmajorind;
CREATE TABLE nowomenmajorind AS
SELECT binpplbyindustry.majorgroup, sum(binaryval)/count(binaryval) AS percentage
FROM binpplbyindustry GROUP BY binpplbyindustry.majorgroup
ORDER BY percentage DESC;
--6
/*highest is biotechnology, with 11% of company founders being women. lowest is
semiconductors/other electric, with 4%*/
 
DROP TABLE nowomenminorind;
CREATE TABLE nowomenminorind AS
SELECT binpplbyindustry.minorgroup, sum(binaryval)/count(binaryval) AS percentage
FROM binpplbyindustry GROUP BY binpplbyindustry.minorgroup
ORDER BY percentage DESC;
--10
/*consumer related higher than biotechnology at 11.9% vs 11.5%, with lowest still
semiconductors/other electric at 4%)*/
 
DROP TABLE nowomensub1;
CREATE TABLE nowomensub1 AS
SELECT binpplbyindustry.subgroup1, sum(binaryval)/count(binaryval) AS percentage
FROM binpplbyindustry GROUP BY binpplbyindustry.subgroup1
ORDER BY percentage DESC;
--18
/*pretty similar to minorind; not much difference/nothing revolutionary to add*/
 
DROP TABLE nowomensub2;
CREATE TABLE nowomensub2 AS
SELECT binpplbyindustry.subgroup2, sum(binaryval)/count(binaryval) AS percentage
FROM binpplbyindustry GROUP BY binpplbyindustry.subgroup2
ORDER BY percentage DESC;
--69
/*consumer products = 15%, consumer services = 14%, energy enhanced recovery = 14%,
med/health services 13%, then biotech research at 13%. Lowest: industrial services at
2.5%, idustrial equipment at 2.9%, power supplies at 3.1%, laser related at 3.3%,
optoelectronics at 3.6%*/
 
DROP TABLE nowomensub3;
CREATE TABLE nowomensub3 AS
SELECT binpplbyindustry.subgroup3, sum(binaryval)/count(binaryval) AS percentage
FROM binpplbyindustry GROUP BY binpplbyindustry.subgroup3
ORDER BY percentage ASC;
--560
/*industrial products=50%!, agriculture=33%, ecommerce selling products = 29%,
biotech food enzymes and cultures = 29%, computers and hardware=28%. lowest: there are
straight up 0 in about 75 different things.*/
 
 
/*up next: build table with coname, women ceo&founder (binary 0/1), women in power (0
or 1), #power positions, #women in power, state, year co received first investment,
no rounds, total$invested*/
 
==STATA regression info==
 
name text file "big.do"
press play - will run in stata
 
 
commands in stata:
summarize numrounds -> gives mean, skewness, kurtosis
summarize numrounds, det -> same as above except w higher level of detail
tab numrounds -> frequency count
gen -> generate a variable
gen invested = totalinvested*1000 = new variable
gen abovemeaninv =0 if totalinvested <. (where . is equal to positive infinity, it's what stata puts in when there's a missing value)
replace abovemeaninv=1 if totalivnested > mean & totalinvested <.
 
reg some measure of performance on some measure of women
 
reg numrounds womeninpower
 
 
if 2(std. err.) < coef, you're good
The relationship between numrounds and womeninpower is highly statistically significant
- R-squared - says of all the variation in the number of rounds, women in power only explains less than one percent of that variation 0.66%
 
 
if you use .log you can automatically replace it whenever you run the script
 
 
rebuild data set to contain men
need year variable - extract year from date
 
 
 
 
multivariate regression
y should be perfromance measure
x should be the explanatory variables
if you run multiple variables with , robust it will be fine
 
i.year = indicator that takes value 1 if it's the year it's invested
=Spring 2016=
[[Category: McNair Projects]]
</includeonly><!-- flush --><!-- flush --><!-- flush -->
 
[[Category: Internal]]
[[Internal Classification: Issue Brief| ]]

Navigation menu