[Script Info]
Title: 
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
Dialogue: 0,0:00:01.42,0:00:05.06,Default,,0000,0000,0000,,- [Narrator] On his quest \Nto master econometrics,
Dialogue: 0,0:00:05.22,0:00:08.81,Default,,0000,0000,0000,,Grasshopper Kamal has \Nmade great progress,
Dialogue: 0,0:00:08.81,0:00:13.66,Default,,0000,0000,0000,,stretching his capabilities \Nand outsmarting his foes.
Dialogue: 0,0:00:14.22,0:00:16.64,Default,,0000,0000,0000,,Alas, today he's despondent,
Dialogue: 0,0:00:16.64,0:00:19.34,Default,,0000,0000,0000,,for one challenge remains unmet.
Dialogue: 0,0:00:19.34,0:00:24.13,Default,,0000,0000,0000,,Kamal cannot yet decode \Nthe scriptures of academic research,
Dialogue: 0,0:00:24.13,0:00:27.35,Default,,0000,0000,0000,,journals like \N"The American Economic Review"
Dialogue: 0,0:00:27.35,0:00:29.08,Default,,0000,0000,0000,,and "Econometrica."
Dialogue: 0,0:00:29.20,0:00:33.50,Default,,0000,0000,0000,,These seemed to him to be inscribed \Nin an obscure foreign tongue.
Dialogue: 0,0:00:33.50,0:00:35.48,Default,,0000,0000,0000,,- [Kamal] Ugh, what the... ?
Dialogue: 0,0:00:36.71,0:00:40.02,Default,,0000,0000,0000,,- These volumes are \Nopaque to the novice, Kamal,
Dialogue: 0,0:00:40.02,0:00:42.20,Default,,0000,0000,0000,,but can be deciphered with study.
Dialogue: 0,0:00:42.47,0:00:45.11,Default,,0000,0000,0000,,Let us learn to read them together.
Dialogue: 0,0:00:52.48,0:00:55.32,Default,,0000,0000,0000,,Let's dive into the West Point study,
Dialogue: 0,0:00:55.32,0:00:58.28,Default,,0000,0000,0000,,published in the "Economics \Nof Education Review."
Dialogue: 0,0:00:58.54,0:01:01.69,Default,,0000,0000,0000,,This paper reports \Non a randomized evaluation
Dialogue: 0,0:01:01.69,0:01:05.86,Default,,0000,0000,0000,,of student electronics use \Nin Economics 101 classrooms.
Dialogue: 0,0:01:06.24,0:01:09.19,Default,,0000,0000,0000,,First, a quick review \Nof the research design.
Dialogue: 0,0:01:09.42,0:01:10.52,Default,,0000,0000,0000,,- Okay.
Dialogue: 0,0:01:11.55,0:01:13.63,Default,,0000,0000,0000,,- [Josh] 'Metrics masters \Nteaching at West Point,
Dialogue: 0,0:01:13.63,0:01:16.62,Default,,0000,0000,0000,,the military college that trains \NAmerican Army officers
Dialogue: 0,0:01:16.62,0:01:19.85,Default,,0000,0000,0000,,designed a randomized trial\Nto answer this question.
Dialogue: 0,0:01:20.37,0:01:23.23,Default,,0000,0000,0000,,These masters randomly assigned\NWest Point cadets
Dialogue: 0,0:01:23.23,0:01:26.38,Default,,0000,0000,0000,,into Economics classes\Noperating under different rules.
Dialogue: 0,0:01:26.60,0:01:28.96,Default,,0000,0000,0000,,Unlike most American colleges,
Dialogue: 0,0:01:28.96,0:01:31.94,Default,,0000,0000,0000,,the West Point default \Nis no electronics.
Dialogue: 0,0:01:32.34,0:01:35.43,Default,,0000,0000,0000,,For purposes of this experiment,\Nsome students were left
Dialogue: 0,0:01:35.43,0:01:38.68,Default,,0000,0000,0000,,in such traditional \Ntechnology-free classes,
Dialogue: 0,0:01:38.68,0:01:41.91,Default,,0000,0000,0000,,no laptops, no tablets \Nand no phones!
Dialogue: 0,0:01:41.91,0:01:43.28,Default,,0000,0000,0000,,[voice echoes]
Dialogue: 0,0:01:43.33,0:01:45.74,Default,,0000,0000,0000,,This is the control group,\Nor baseline case.
Dialogue: 0,0:01:46.21,0:01:49.20,Default,,0000,0000,0000,,Another group was allowed\Nto use electronics.
Dialogue: 0,0:01:49.27,0:01:52.70,Default,,0000,0000,0000,,This is the treatment group,\Nsubject to a changed environment.
Dialogue: 0,0:01:53.31,0:01:55.86,Default,,0000,0000,0000,,The treatment in this case\Nis the unrestricted use
Dialogue: 0,0:01:55.86,0:01:58.11,Default,,0000,0000,0000,,of laptops or tablets in class.
Dialogue: 0,0:01:58.84,0:02:01.81,Default,,0000,0000,0000,,Every causal question \Nhas a clear outcome,
Dialogue: 0,0:02:01.86,0:02:05.28,Default,,0000,0000,0000,,the variables we hope to influence\Ndefined in advance of the study.
Dialogue: 0,0:02:05.86,0:02:08.38,Default,,0000,0000,0000,,The outcomes in the West Point\Nelectronics study
Dialogue: 0,0:02:08.38,0:02:09.99,Default,,0000,0000,0000,,are final exam scores.
Dialogue: 0,0:02:10.05,0:02:13.36,Default,,0000,0000,0000,,The study seeks to answer\Nthe following question,
Dialogue: 0,0:02:13.36,0:02:17.30,Default,,0000,0000,0000,,what is the causal effect \Nof classroom electronics on learning
Dialogue: 0,0:02:17.30,0:02:19.76,Default,,0000,0000,0000,,as measured by exam scores?
Dialogue: 0,0:02:20.63,0:02:24.20,Default,,0000,0000,0000,,- Economics journal articles \Nusually begin with a table
Dialogue: 0,0:02:24.20,0:02:26.93,Default,,0000,0000,0000,,of descriptive statistics,\Ngiving key facts
Dialogue: 0,0:02:26.93,0:02:28.50,Default,,0000,0000,0000,,about the study sample.
Dialogue: 0,0:02:28.50,0:02:31.78,Default,,0000,0000,0000,,- Oh my gosh, I remember this table,\Nso confusing!
Dialogue: 0,0:02:31.78,0:02:36.67,Default,,0000,0000,0000,,- [Narrator] Columns 1 to 3 report \Nmean, or average, characteristics.
Dialogue: 0,0:02:36.74,0:02:39.69,Default,,0000,0000,0000,,These give a sense \Nof who we're studying.
Dialogue: 0,0:02:39.95,0:02:43.74,Default,,0000,0000,0000,,Let's start with column 1 \Nwhich describes covariates
Dialogue: 0,0:02:43.74,0:02:45.25,Default,,0000,0000,0000,,in the control group.
Dialogue: 0,0:02:45.41,0:02:48.97,Default,,0000,0000,0000,,Covariates are characteristics \Nof the control and treatment groups
Dialogue: 0,0:02:48.97,0:02:51.62,Default,,0000,0000,0000,,measured before \Nthe experiment begins.
Dialogue: 0,0:02:51.62,0:02:56.99,Default,,0000,0000,0000,,For example, we see the control group\Nhas an average age a bit over 20.
Dialogue: 0,0:02:57.32,0:03:00.34,Default,,0000,0000,0000,,Many of these covariates \Nare dummy variables.
Dialogue: 0,0:03:00.79,0:03:05.67,Default,,0000,0000,0000,,A dummy variable can only have \Ntwo values, a zero or a one.
Dialogue: 0,0:03:05.98,0:03:10.02,Default,,0000,0000,0000,,For example, student gender\Nis captured by a dummy variable
Dialogue: 0,0:03:10.02,0:03:13.15,Default,,0000,0000,0000,,that equals one for women\Nand zero for men.
Dialogue: 0,0:03:13.25,0:03:16.58,Default,,0000,0000,0000,,The mean of this variable \Nis the proportion female.
Dialogue: 0,0:03:16.82,0:03:20.65,Default,,0000,0000,0000,,We also see that the control group \Nis 13% Hispanic
Dialogue: 0,0:03:20.65,0:03:23.77,Default,,0000,0000,0000,,and 19% had prior military service.
Dialogue: 0,0:03:25.04,0:03:26.64,Default,,0000,0000,0000,,The table notes are key.
Dialogue: 0,0:03:26.64,0:03:28.69,Default,,0000,0000,0000,,Refer to these \Nas you scan the table.
Dialogue: 0,0:03:29.10,0:03:33.37,Default,,0000,0000,0000,,These notes explain what's shown \Nin each column and panel.
Dialogue: 0,0:03:39.48,0:03:43.38,Default,,0000,0000,0000,,The notes tell us, for example,\Nthat standard deviations
Dialogue: 0,0:03:43.38,0:03:45.18,Default,,0000,0000,0000,,are reported in brackets.
Dialogue: 0,0:03:45.95,0:03:49.60,Default,,0000,0000,0000,,Standard deviations tell us how\Nspread out the data are.
Dialogue: 0,0:03:50.28,0:03:54.89,Default,,0000,0000,0000,,For example, a standard deviation \Nof 0.52 tells us that most
Dialogue: 0,0:03:54.89,0:03:59.23,Default,,0000,0000,0000,,of the control group's GPAs\Nfall between 2.35,
Dialogue: 0,0:03:59.23,0:04:03.45,Default,,0000,0000,0000,,which is 0.52 below \Nthe mean GPA of 2.87,
Dialogue: 0,0:04:03.45,0:04:08.34,Default,,0000,0000,0000,,and 3.39, which is 0.52 above 2.87.
Dialogue: 0,0:04:09.00,0:04:12.22,Default,,0000,0000,0000,,A lower standard deviation\Nwould mean the GPAs were
Dialogue: 0,0:04:12.22,0:04:14.40,Default,,0000,0000,0000,,more tightly clustered \Naround the mean.
Dialogue: 0,0:04:14.55,0:04:17.45,Default,,0000,0000,0000,,- [Kamal] Yeah, but they're missing\Nfor most of the variables.
Dialogue: 0,0:04:17.50,0:04:18.60,Default,,0000,0000,0000,,- [Narrator] That's right.
Dialogue: 0,0:04:18.60,0:04:22.50,Default,,0000,0000,0000,,Masters usually omit \Nstandard deviations for dummies
Dialogue: 0,0:04:22.50,0:04:26.50,Default,,0000,0000,0000,,because the mean of this variable\Ndetermines its standard deviation.
Dialogue: 0,0:04:26.94,0:04:31.37,Default,,0000,0000,0000,,This study compares two treatment \Ngroups with the control group.
Dialogue: 0,0:04:31.37,0:04:35.75,Default,,0000,0000,0000,,The first was allowed free use \Nof laptops and tablets.
Dialogue: 0,0:04:35.75,0:04:38.25,Default,,0000,0000,0000,,The second treatment\Nwas more restrictive,
Dialogue: 0,0:04:38.25,0:04:41.55,Default,,0000,0000,0000,,allowing only tablets placed \Nflat on the desk.
Dialogue: 0,0:04:42.15,0:04:45.24,Default,,0000,0000,0000,,The treatment groups\Nlook much like the control group.
Dialogue: 0,0:04:46.31,0:04:51.27,Default,,0000,0000,0000,,This takes us to the next feature \Nof this table, columns 4 through 6
Dialogue: 0,0:04:51.41,0:04:54.56,Default,,0000,0000,0000,,use statistical tests to compare\Nthe characteristics
Dialogue: 0,0:04:54.56,0:04:57.59,Default,,0000,0000,0000,,of the treatment and control group\Nbefore the experiment.
Dialogue: 0,0:04:58.02,0:05:01.67,Default,,0000,0000,0000,,In column 4, the two treatment \Ngroups are combined.
Dialogue: 0,0:05:01.86,0:05:04.84,Default,,0000,0000,0000,,You can see that the difference\Nin proportion female
Dialogue: 0,0:05:04.84,0:05:09.69,Default,,0000,0000,0000,,between the treatment \Nand control group is only 0.03.
Dialogue: 0,0:05:09.99,0:05:13.74,Default,,0000,0000,0000,,The difference is not \Nstatistically significant.
Dialogue: 0,0:05:14.29,0:05:17.20,Default,,0000,0000,0000,,It is the sort of difference\Nwe can easily put down
Dialogue: 0,0:05:17.20,0:05:20.50,Default,,0000,0000,0000,,to chance results \Nin our sample selection process.
Dialogue: 0,0:05:20.50,0:05:22.13,Default,,0000,0000,0000,,- [Kamal] Hmm, how do we know that?
Dialogue: 0,0:05:22.13,0:05:23.79,Default,,0000,0000,0000,,- [Narrator] Remember \Nthe rule of thumb?
Dialogue: 0,0:05:23.79,0:05:26.97,Default,,0000,0000,0000,,Statistical estimates \Nthat exceed the standard error
Dialogue: 0,0:05:26.97,0:05:29.88,Default,,0000,0000,0000,,by a multiple of 2 \Nin absolute value
Dialogue: 0,0:05:30.10,0:05:33.100,Default,,0000,0000,0000,,are usually said\Nto be statistically significant.
Dialogue: 0,0:05:35.13,0:05:38.42,Default,,0000,0000,0000,,The standard error is 0.03,
Dialogue: 0,0:05:38.57,0:05:41.48,Default,,0000,0000,0000,,same as the difference \Nin proportion female.
Dialogue: 0,0:05:42.02,0:05:46.04,Default,,0000,0000,0000,,So the ratio of the latter \Nto the former is only 1,
Dialogue: 0,0:05:46.04,0:05:48.61,Default,,0000,0000,0000,,which of course is less than 2.
Dialogue: 0,0:05:48.61,0:05:51.19,Default,,0000,0000,0000,,- [Kamal] Uh huh. So none\Nof the treatment/control differences
Dialogue: 0,0:05:51.19,0:05:54.33,Default,,0000,0000,0000,,in the table are more than twice\Ntheir standard errors.
Dialogue: 0,0:05:54.33,0:05:55.79,Default,,0000,0000,0000,,- [Narrator] Correct.
Dialogue: 0,0:05:55.79,0:05:59.00,Default,,0000,0000,0000,,The random division of students\Nappears to have succeeded
Dialogue: 0,0:05:59.01,0:06:01.94,Default,,0000,0000,0000,,in creating groups \Nthat are indeed comparable.
Dialogue: 0,0:06:02.85,0:06:06.36,Default,,0000,0000,0000,,We can be confident therefore\Nthat any later differences
Dialogue: 0,0:06:06.36,0:06:09.83,Default,,0000,0000,0000,,in classroom achievement\Nare the result of the experimental
Dialogue: 0,0:06:09.83,0:06:12.58,Default,,0000,0000,0000,,intervention rather \Nthan a reflection
Dialogue: 0,0:06:12.58,0:06:14.65,Default,,0000,0000,0000,,of preexisting differences.
Dialogue: 0,0:06:14.65,0:06:17.23,Default,,0000,0000,0000,,Ceteris paribus achieved!
Dialogue: 0,0:06:17.36,0:06:20.72,Default,,0000,0000,0000,,- [Kamal] Cool. Wait, \Nwhat about the bottom,
Dialogue: 0,0:06:20.72,0:06:22.52,Default,,0000,0000,0000,,the numbers with the stars?
Dialogue: 0,0:06:22.71,0:06:25.48,Default,,0000,0000,0000,,Those differences are a lot more\Nthan double the standard error.
Dialogue: 0,0:06:25.48,0:06:27.40,Default,,0000,0000,0000,,- [Narrator] Good eye, Kamal!
Dialogue: 0,0:06:27.40,0:06:29.39,Default,,0000,0000,0000,,The table has many numbers.
Dialogue: 0,0:06:29.39,0:06:32.25,Default,,0000,0000,0000,,Those in Panel B are important too.
Dialogue: 0,0:06:32.25,0:06:35.72,Default,,0000,0000,0000,,This panel measures the extent \Nto which students in treatment
Dialogue: 0,0:06:35.72,0:06:39.05,Default,,0000,0000,0000,,and control groups actually use \Ncomputers in class.
Dialogue: 0,0:06:39.50,0:06:42.75,Default,,0000,0000,0000,,The treatment here was \Nto allow computer use.
Dialogue: 0,0:06:42.75,0:06:46.07,Default,,0000,0000,0000,,The researchers must show\Nthat students allowed
Dialogue: 0,0:06:46.08,0:06:49.45,Default,,0000,0000,0000,,to use computers took advantage \Nof the opportunity to do so.
Dialogue: 0,0:06:49.90,0:06:53.03,Default,,0000,0000,0000,,If they didn't, then there's \Nreally no treatment.
Dialogue: 0,0:06:53.25,0:06:57.80,Default,,0000,0000,0000,,Luckily, 81% of those \Nin the first treatment group
Dialogue: 0,0:06:57.80,0:07:01.83,Default,,0000,0000,0000,,used computers compared \Nwith none in the control group.
Dialogue: 0,0:07:02.08,0:07:05.03,Default,,0000,0000,0000,,And many in the second \Ntablet treatment group
Dialogue: 0,0:07:05.03,0:07:06.100,Default,,0000,0000,0000,,used computers as well.
Dialogue: 0,0:07:07.21,0:07:09.73,Default,,0000,0000,0000,,These differences \Nin computer use are large
Dialogue: 0,0:07:09.73,0:07:11.80,Default,,0000,0000,0000,,and statistically significant.
Dialogue: 0,0:07:12.08,0:07:15.37,Default,,0000,0000,0000,,We also get to see \Nthe sample size in each group.
Dialogue: 0,0:07:15.37,0:07:18.10,Default,,0000,0000,0000,,- [Kamal] The stars \Nare just like decoration?
Dialogue: 0,0:07:18.10,0:07:21.75,Default,,0000,0000,0000,,- [Narrator] Some academic papers\Nuse stars to indicate differences
Dialogue: 0,0:07:21.75,0:07:23.86,Default,,0000,0000,0000,,that are statistically significant.
Dialogue: 0,0:07:23.86,0:07:26.92,Default,,0000,0000,0000,,This makes them jump out at you.
Dialogue: 0,0:07:26.92,0:07:31.62,Default,,0000,0000,0000,,Here three stars indicate that \Nthe result is statistically different
Dialogue: 0,0:07:31.62,0:07:34.94,Default,,0000,0000,0000,,from zero with a p value \Nless than 1%.
Dialogue: 0,0:07:35.34,0:07:39.44,Default,,0000,0000,0000,,In other words, there's less \Nthan a 1 in 100 chance
Dialogue: 0,0:07:39.44,0:07:42.17,Default,,0000,0000,0000,,this result is purely \Na chance finding.
Dialogue: 0,0:07:42.17,0:07:43.18,Default,,0000,0000,0000,,[applause]
Dialogue: 0,0:07:43.18,0:07:48.60,Default,,0000,0000,0000,,Two stars indicate a 1 in 20\Nor 5% chance of a chance finding.
Dialogue: 0,0:07:48.80,0:07:53.20,Default,,0000,0000,0000,,And one star denotes results \Nwe might see as often as 10%
Dialogue: 0,0:07:53.20,0:07:56.04,Default,,0000,0000,0000,,of the time merely due to chance.
Dialogue: 0,0:07:56.47,0:07:59.74,Default,,0000,0000,0000,,Today, stars are seen \Nas a little old fashioned.
Dialogue: 0,0:07:59.74,0:08:01.44,Default,,0000,0000,0000,,Some journals omit them.
Dialogue: 0,0:08:01.44,0:08:03.32,Default,,0000,0000,0000,,- [Kamal] What about \Nthose last two columns?
Dialogue: 0,0:08:03.32,0:08:06.01,Default,,0000,0000,0000,,- [Narrator] Unlike column 4,\Nwhich combines
Dialogue: 0,0:08:06.01,0:08:09.69,Default,,0000,0000,0000,,both treatment groups into one, \Nthese last two columns
Dialogue: 0,0:08:09.69,0:08:12.36,Default,,0000,0000,0000,,look separately\Nat treatment/control differences
Dialogue: 0,0:08:12.36,0:08:14.36,Default,,0000,0000,0000,,for each treatment group.
Dialogue: 0,0:08:14.36,0:08:17.44,Default,,0000,0000,0000,,This provides a more detailed \Nanalysis of balance.
Dialogue: 0,0:08:17.86,0:08:21.06,Default,,0000,0000,0000,,Also, for now, \Nyou can ignore this row
Dialogue: 0,0:08:21.06,0:08:24.23,Default,,0000,0000,0000,,which provides \Nanother test of significance.
Dialogue: 0,0:08:24.23,0:08:28.92,Default,,0000,0000,0000,,Now we get to the article's \Npunchline, table 4.
Dialogue: 0,0:08:29.93,0:08:32.99,Default,,0000,0000,0000,,This table reports\Nregression estimates
Dialogue: 0,0:08:32.99,0:08:36.98,Default,,0000,0000,0000,,of the effects of electronics use\Non measures of student learning.
Dialogue: 0,0:08:37.17,0:08:40.03,Default,,0000,0000,0000,,- [Kamal Why does the study \Nreport regression estimates?
Dialogue: 0,0:08:40.03,0:08:42.20,Default,,0000,0000,0000,,See, that's why I was getting lost.
Dialogue: 0,0:08:42.20,0:08:44.81,Default,,0000,0000,0000,,I thought one reason \Nwhy we liked randomized trials
Dialogue: 0,0:08:44.81,0:08:47.26,Default,,0000,0000,0000,,is that we use them \Nto obtain causal effects
Dialogue: 0,0:08:47.26,0:08:50.14,Default,,0000,0000,0000,,simply by comparing \Ntreatment and control groups.
Dialogue: 0,0:08:50.29,0:08:53.49,Default,,0000,0000,0000,,Since these groups are balanced,\Nno need to use regression.
Dialogue: 0,0:08:53.49,0:08:55.29,Default,,0000,0000,0000,,- [Narrator] Well said, Kamal.
Dialogue: 0,0:08:55.29,0:08:59.27,Default,,0000,0000,0000,,In practice, it's customary \Nto report regression estimates
Dialogue: 0,0:08:59.27,0:09:00.84,Default,,0000,0000,0000,,for two reasons.
Dialogue: 0,0:09:00.98,0:09:05.01,Default,,0000,0000,0000,,First, evidence of balance \Nnot withstanding, an abundance
Dialogue: 0,0:09:05.01,0:09:09.06,Default,,0000,0000,0000,,of caution might lead the analyst\Nto allow for chance differences.
Dialogue: 0,0:09:09.39,0:09:13.36,Default,,0000,0000,0000,,Second, regression estimates\Nare likely to be more precise.
Dialogue: 0,0:09:13.73,0:09:16.51,Default,,0000,0000,0000,,That is, they have lower \Nstandard errors than
Dialogue: 0,0:09:16.51,0:09:18.89,Default,,0000,0000,0000,,the simple treatment \Ncontrol comparisons.
Dialogue: 0,0:09:19.90,0:09:22.53,Default,,0000,0000,0000,,The dependent variable \Nin this study
Dialogue: 0,0:09:22.53,0:09:24.30,Default,,0000,0000,0000,,is the outcome of interest.
Dialogue: 0,0:09:24.65,0:09:27.72,Default,,0000,0000,0000,,Since the question at hand\Nis how classroom electronics
Dialogue: 0,0:09:27.72,0:09:32.62,Default,,0000,0000,0000,,affect learning, a good outcome\Nis the Economics final exam score.
Dialogue: 0,0:09:32.92,0:09:37.17,Default,,0000,0000,0000,,Each column reports results\Nfrom a different regression model.
Dialogue: 0,0:09:37.30,0:09:40.48,Default,,0000,0000,0000,,Models are distinguished \Nby the control variables
Dialogue: 0,0:09:40.48,0:09:44.71,Default,,0000,0000,0000,,or covariates they include\Nbesides treatment status.
Dialogue: 0,0:09:44.71,0:09:48.42,Default,,0000,0000,0000,,Estimates with no covariates \Nare simple comparisons
Dialogue: 0,0:09:48.42,0:09:50.50,Default,,0000,0000,0000,,of treatment and control groups.
Dialogue: 0,0:09:50.62,0:09:52.72,Default,,0000,0000,0000,,- [Kamal] I thought \Nthey just forgot to fill it out.
Dialogue: 0,0:09:52.72,0:09:56.67,Default,,0000,0000,0000,,- [Narrator] Column 1 suggests\Nelectronics use reduced
Dialogue: 0,0:09:56.67,0:10:00.84,Default,,0000,0000,0000,,final exam scores\Nby 0.28 standard deviations.
Dialogue: 0,0:10:01.10,0:10:06.55,Default,,0000,0000,0000,,In our last lesson, Master Joshway\Nexplained, we use standard deviation
Dialogue: 0,0:10:06.55,0:10:10.50,Default,,0000,0000,0000,,units because these units \Nare easily compared across studies.
Dialogue: 0,0:10:11.00,0:10:13.70,Default,,0000,0000,0000,,Column 2 reports results \Nfrom a model
Dialogue: 0,0:10:13.70,0:10:15.95,Default,,0000,0000,0000,,that adds demographic controls.
Dialogue: 0,0:10:15.95,0:10:19.50,Default,,0000,0000,0000,,Here we're comparing test scores\Nbut holding constant factors
Dialogue: 0,0:10:19.50,0:10:21.44,Default,,0000,0000,0000,,such as age and sex.
Dialogue: 0,0:10:21.89,0:10:25.28,Default,,0000,0000,0000,,Column 3 reports results\Nfrom a model that adds GPA
Dialogue: 0,0:10:25.28,0:10:27.19,Default,,0000,0000,0000,,to the list of covariates.
Dialogue: 0,0:10:27.60,0:10:30.50,Default,,0000,0000,0000,,Column 4 adds ACT scores.
Dialogue: 0,0:10:30.50,0:10:33.50,Default,,0000,0000,0000,,Analysts often report \Nresults this way,
Dialogue: 0,0:10:33.50,0:10:36.80,Default,,0000,0000,0000,,starting with models that include \Nfew or no covariates
Dialogue: 0,0:10:36.80,0:10:40.45,Default,,0000,0000,0000,,and then reporting estimates \Nfrom models that add more
Dialogue: 0,0:10:40.45,0:10:43.59,Default,,0000,0000,0000,,and more covariates \Nas we move across columns.
Dialogue: 0,0:10:44.04,0:10:46.80,Default,,0000,0000,0000,,Looking across columns,\Nwhat do you notice?
Dialogue: 0,0:10:47.25,0:10:49.92,Default,,0000,0000,0000,,- [Kamal] Well, the coefficient \Non using a computer is always
Dialogue: 0,0:10:49.92,0:10:51.64,Default,,0000,0000,0000,,a pretty big negative number.
Dialogue: 0,0:10:51.64,0:10:53.00,Default,,0000,0000,0000,,- [Narrator] That's right!
Dialogue: 0,0:10:53.00,0:10:56.46,Default,,0000,0000,0000,,We can also see that \Nthe standard errors are small enough
Dialogue: 0,0:10:56.46,0:11:00.20,Default,,0000,0000,0000,,to make these negative results\Nstatistically significant.
Dialogue: 0,0:11:00.32,0:11:04.20,Default,,0000,0000,0000,,In other words, the primary \Ntakeaway from this experiment
Dialogue: 0,0:11:04.20,0:11:08.32,Default,,0000,0000,0000,,is that electronics in the classroom\Nreduce student learning.
Dialogue: 0,0:11:08.77,0:11:11.88,Default,,0000,0000,0000,,- [Kama] GPA and ACT scores \Nare also significant.
Dialogue: 0,0:11:11.88,0:11:13.60,Default,,0000,0000,0000,,Why is that?
Dialogue: 0,0:11:13.60,0:11:15.10,Default,,0000,0000,0000,,- [Narrator] Good observation!
Dialogue: 0,0:11:15.10,0:11:16.87,Default,,0000,0000,0000,,That's not surprising.
Dialogue: 0,0:11:16.87,0:11:20.27,Default,,0000,0000,0000,,We expect these variables \Nto predict college performance.
Dialogue: 0,0:11:20.27,0:11:21.98,Default,,0000,0000,0000,,- [Kamal] Oh right, of course.
Dialogue: 0,0:11:21.98,0:11:24.82,Default,,0000,0000,0000,,Kids who got better grades before\Nare more likely to get
Dialogue: 0,0:11:24.82,0:11:26.32,Default,,0000,0000,0000,,a better grade in this course.
Dialogue: 0,0:11:26.32,0:11:29.85,Default,,0000,0000,0000,,- [Narrator] You'll also notice a lot \Nof other information on this table.
Dialogue: 0,0:11:29.85,0:11:34.23,Default,,0000,0000,0000,,Remaining panels in the table\Nreport effects of electronics use
Dialogue: 0,0:11:34.23,0:11:36.93,Default,,0000,0000,0000,,on components of the final exam,
Dialogue: 0,0:11:36.93,0:11:39.47,Default,,0000,0000,0000,,such as the multiple \Nchoice questions.
Dialogue: 0,0:11:39.47,0:11:43.28,Default,,0000,0000,0000,,These results are mostly consistent \Nwith computer use effects
Dialogue: 0,0:11:43.28,0:11:45.22,Default,,0000,0000,0000,,on overall scores.
Dialogue: 0,0:11:45.22,0:11:47.74,Default,,0000,0000,0000,,- [Kamal] What about the rows\Nnot in English?
Dialogue: 0,0:11:47.74,0:11:50.56,Default,,0000,0000,0000,,- [Narrator] These rows give \Nadditional statistical information.
Dialogue: 0,0:11:50.83,0:11:53.98,Default,,0000,0000,0000,,R-squared is a measure\Nof goodness of fit.
Dialogue: 0,0:11:54.31,0:11:58.01,Default,,0000,0000,0000,,This isn't too important, though \Nsome readers may want to know it.
Dialogue: 0,0:11:58.66,0:12:02.83,Default,,0000,0000,0000,,Other rows report on alternative \Ntests of statistical significance
Dialogue: 0,0:12:02.83,0:12:05.03,Default,,0000,0000,0000,,that you can ignore for now.
Dialogue: 0,0:12:05.03,0:12:07.64,Default,,0000,0000,0000,,- [Kamal] Oh my gosh,\Nthese tables aren't that hard!
Dialogue: 0,0:12:07.64,0:12:09.49,Default,,0000,0000,0000,,Thank you so much.
Dialogue: 0,0:12:09.49,0:12:11.79,Default,,0000,0000,0000,,- [Narrator] Next up is regression.
Dialogue: 0,0:12:11.79,0:12:13.18,Default,,0000,0000,0000,,See you then!
Dialogue: 0,0:12:15.97,0:12:17.26,Default,,0000,0000,0000,,♪ [music] ♪
Dialogue: 0,0:12:17.26,0:12:20.58,Default,,0000,0000,0000,,You're on your way \Nto mastering econometrics.
Dialogue: 0,0:12:20.83,0:12:22.78,Default,,0000,0000,0000,,Make sure this video sticks
Dialogue: 0,0:12:22.78,0:12:24.98,Default,,0000,0000,0000,,by taking a few \Nquick practice questions.
Dialogue: 0,0:12:25.15,0:12:28.86,Default,,0000,0000,0000,,Or, if you're ready,\Nclick for the next video.
Dialogue: 0,0:12:28.86,0:12:32.62,Default,,0000,0000,0000,,You can also check out MRU's \Nwebsite for more courses,
Dialogue: 0,0:12:32.62,0:12:35.30,Default,,0000,0000,0000,,teacher resources and more.