# Political Science

MEAN RANGE STDDEV/FORMAT=NOTABLES.

Explanation:

Once the ANES2016REVISED data set is downloaded, we then apply the PW2016_FULL weight. • This adjusts for the unintentional oversampling of wealthy, highly educated, and female

respondents by reducing the impact of their responses. The next syntax line produces a frequency distribution for each of those variables, along with the requested statistics for each.

• SPSS does not directly calculate the MAD • The ‘FORMAT’ command eliminates the listing of each of the potential 101 categories of the

feeling thermometer (0-100), and just produces the requested statistics.

TASK 1:

Interpret each of the statistical values as you did in Stat 1. Just discuss the standard deviation

as a weighted (i.e., extreme cases count more heavily) MAD.

TASK 2:

Compare the four statistics across each of the two variables. What do they tell you about the

distribution of feelings towards the two major-party candidates?

Frequencies

Statistics

PRE: Feeling

Thermometer:

Democratic

Presidential cand

PRE: Feeling

Thermometer:

Republican

Presidential cand

N Valid 3598 3597

Missing 51 52

Mean 42.16 36.26

Median 40.00 30.00

Std. Deviation 34.099 34.867

Range 100 100

‘Missing’ is the number of respondents (V16=51/V17=v52) who didn’t answer the associated

question.

MEANS TABLES=V16 V17 BY V22A_30D/CELLS=MEDIAN MEAN RANGE STDDEV COUNT. MEANS TABLES=V16 V17 BY V42E_30D/CELLS=MEDIAN MEAN RANGE STDDEV COUNT. MEANS TABLES=V16 V17 BY V47_30D/CELLS=MEDIAN MEAN RANGE STDDEV COUNT. Explanation: The MEANS procedure makes it easy for us to calculate statistics for a variable or variables for different subgroups within our sample. The first listed syntax repeats our request from Tasks 1 and 2 but creates a different set of statistics for each of the two categories of V22A_30D (Do you favor or oppose the ACA/Obamacare?). ‘Count’ is the number of cases (N) in each category.

• To make your life easy, I have dichotomized (converted into only two categories) V22A into just the two categories listed:

o 1=Favor o 2=Oppose

• I have done the same for opinions about federal environmental funding (V42E_30D):

o 1=Increase funding o 2=Decrease funding or keep the same o

• and gun ownership (V47_30D): o 1=Make more difficult o 2=Make easier or keep laws the same

TASK 3:

Choose only one (1) of the dichotomized variables (V22A_30D, V42E_30D, V47_30D).

Compare the statistics between each category of your chosen dichotomized variables.

• You should realize by now that the RANGE is not very useful as we will always have respondents who give a ‘0’ evaluation as well as a ‘100’. You may ignore that statistic.

• You should have one set of comparisons for V16 and a separate one for V17. • So, for example, you should compare the median, mean, and standard deviation of V16

between those who favor Obamacare (1) and those who oppose it (2). You then do the same for the statistics for V17.

• Be as detailed as possible in your comparison. More is better than less.

Here are the relevant tables produced by each MEANS procedure: MEANS TABLES=V16 V17 BY V22A_30D/CELLS=MEDIAN MEAN RANGE STDDEV COUNT.

Means

Report

ACA

PRE: Feeling

Thermometer:

Democratic

Presidential cand

PRE: Feeling

Thermometer:

Republican

Presidential cand

Favor Median 70.00 .00

Mean 65.01 16.50

Range 100 100

Std. Deviation 27.412 25.879

N 1360 1373

Oppose Median 5.00 61.00

Mean 19.51 56.61

Range 100 100

Std. Deviation 25.543 32.193

N 1470 1462

Total Median 40.00 30.00

Mean 41.37 37.18

Range 100 100

Std. Deviation 34.883 35.503

N 2831 2835

• Notice that the number of respondents (N) is lower than the number in the original tables

(Task 1-2). • Reason? These tables eliminate not only those who didn’t answer V16 or V17, but also those

who didn’t answer or responded ‘Neither favor nor oppose’ to V22A_30D. • The same holds for the other tables printed below, but the drop is minimal as there was no

middle category to remove.

MEANS TABLES=V16 V17 BY V42E_30D/CELLS=MEDIAN MEAN RANGE STDDEV COUNT.

Means

Report

ENVIRONMENT

PRE: Feeling

Thermometer:

Democratic

Presidential cand

PRE: Feeling

Thermometer:

Republican

Presidential cand

Increase Median 60.00 2.00

Mean 54.69 23.91

Range 100 100

Std. Deviation 31.728 30.435

N 1892 1893

Keep the same or decrease Median 15.00 60.00

Mean 28.13 50.17

Range 100 100

Std. Deviation 31.071 34.261

N 1691 1690

Total Median 40.00 30.00

Mean 42.15 36.30

Range 100 100

Std. Deviation 34.099 34.852

N 3583 3582

MEANS TABLES=V16 V17 BY V47_30D/CELLS=MEDIAN MEAN RANGE STDDEV COUNT.

Means

Report

GUNS

PRE: Feeling

Thermometer:

Democratic

Presidential cand

PRE: Feeling

Thermometer:

Republican

Presidential cand

Make more difficult Median 60.00 1.00

Mean 56.57 21.10

Range 100 100

Std. Deviation 31.013 28.252

N 1913 1920

Make easier or keep rules the

same

Median 15.00 60.00

Mean 25.62 53.74

Range 100 100

Std. Deviation 29.778 33.590

N 1668 1660

Total Median 40.00 30.00

Mean 42.15 36.23

Range 100 100

Std. Deviation 34.132 34.871

N 3581 3580