Ethnicity (French, Brazilian, etc).
Phone number (519-884-0710, etc.).
Price ($5, $25.99, etc.).
Computer IP address (12-346-378-90, etc.).
Likely to vote
We should generate a frequency table.
We should compute some basic statistics, such as mean, median, mode.
We should execute a syntax file
We should collapse it into fewer categories.
The survey item is descriptive.
The survey item is reliable.
The survey item has validity.
The survey item avoids the ecological fallacy.
Number of people in a family.
Speed of travel of a car.
Gender of students in a class.
A person’s weight.
The proportion must be squared.
The proportion must be divided by 10.
The proportion must be multiplied by the square root of the sample size.
The proportion must be multiplied by 100.
They provide an indication of the size of a sample.
They provide an indication of the variety of scores within a distribution.
They provide an indication of the typical, or most representative, value in a distribution.
They provide an indication of the adequacy of the selection criteria for the sample.
SELECT CASES q23.
RECODE q23 (1 to 7=keep) (8,9=sysmiss) into q23b.
COMPUTE socmedia = q23 - 8 - 9.
MISSING VALUES q23 (8,9).
The research made a mistake: The variable is continuous and should not be treated as ordinal
The researcher computed the mean properly, although the correct mean is 2.63.
The research made a mistake: The variable is ordinal and should not be treated as continuous.
The researcher was correct, sort of: The variable is nominal and should not be treated as categorical, but in this case, that's probably okay.
In the VALUE LABELS command, the series of labels needs to be separated by a comma.
Misplaced command terminator in the VARIABLE LABELS command.
No central tendency command was included.
In the RECODE command, the subcommand "into" needs to be UPPER CASE.
Standard deviation around the mean.
Histogram and box plots.
Index of Qualitative Variation.
The true population mean may be µ = 50, but a sample of 2100 yields a mean of 49 and a standard deviation of 19. There is a 99% chance that the true population proportion ranges somewhere between 47.9 to 50.1, therefore we can reject the null hypothesis given this low probability of error.
The true population mean may be µ = 50, but a sample of 1300 yields a mean of 48, with a standard deviation of 23. This sample's statistics are improbable, so we reject the hypothesized value of 50 at p<.01.
The true population mean may be µ = 50, but a sample of 5000 shows a mean of 46, which is four whole points away from 50. We therefore can reject the value of 50 at an alpha of .001, which corresponds to 3.3 z scores, much lower than the difference.
The true population mean may be µ = 50, but a sample of 1300 yields a mean of 48, and a standard deviation of 23, which is not improbable. So we do not reject the hypothesized value of 50 at p<.10.
The true population mean may be µ = 50, but a sample of 2100 yields a mean of 49 and a standard deviation of 19. There is a 95% chance that the true population proportion ranges somewhere between 48.2 and 49.8, therefore we cannot reject the null hypothesis as the probability of error is too high.
Using math to study patterns in a way that allows us to derive a precise understanding of phenomena.
Counting and enumerating phenomena so that we can effectively communicate our research.
It assumes that all of reality is quantifiable, and thus, all reality can be explained in quantitative expressions.
It addresses the problem of using samples to generalize about populations.
Assuming a null hypothesis of .50, the calculated z score is 2, which is above 1.96, so reject the null hypothesis with the probability of committing a Type I error no greater than .05.
Cannot infer anything as the sample is too small.
A 99% confidence interval suggests 1% to 59% of university students had voted.
A 95% confidence interval suggests 10.4% to 49.6% of university students had voted.
Only I is correct
Only ii and iii are correct
All of the three choices are correct
Only I and iii are correct
None of the options are correct
You gather a sample of 2100 voters, and it shows the party’s level of support at 36.6%. You test this against the hypothesized 40%, and find that 40% is implausible (p<.001).
You gather a sample of 350 voters, and the mean level of support is 42.5%, which is two-and-a-half points about the target of 40%. With an assumed standard deviation of about .50, this yields a 99% confidence interval ranging from 42.4% to 42.6%.
You gather a sample of 600 voters, and it shows the party’s level of support at 37%. This is three full points below the 40%, so you do not advise calling an election at this time.
You gather a sample of 2900 voters, and it shows the party’s level of support at 37.5%. You test this against the hypothesized 40%, and find that 40% is implausible (p<.05)
The probability of Type 1 and Type II errors has been adequately estimated to ensure that the test is valid.
The difference detected in a hypothesis test is strong enough to have important implications in the real world.
The sample used is large enough to be informative about the general population.
The difference between a sample statistic and population parameter is unlikely to occur by random chance alone.