12/09_MEETING Flashcards
What are four ways of thinking about inferences? (as in to infer information)
- Dichotomous thinking (“You want to find out if Filipinos are more intelligent than your average person, so for that reason you cannot study a population, so you come up with a sample. Observe mean IQ of Filipinos and compare it with typical IQ of an average person. And with that comparison, you make an either/or conclusion, and say it is statistically significant or nonsignificant. It is an either/or. Nothing in between. Assuming it’s statistically significant, you can make a conclusion about the population, e.g. Filipinos are smarter. We’re stuck in dichotomous thinking because this is the simplest—very childlike.)
- Estimation thinking (Not an either/or—give us the value! Estimate the actual value. Two types of estimates—point estimates and interval estimates. Point estimates are single values, such as the measures of central tendency, or variability. So when you say the average IQ of Filipinos is 125, that’s a point estimate. When it comes to p-value, the observation of this happening based on a certain theory is 0.06, which is again, a point estimate. We wanna get away from these single point estimates. We look at interval estimates that have an upper and a lower limit. So, where are we right now? The teaching of statistics is misunderstood right now as dichotomous thinking, but we want to get away from that right now.
- Meta-analytic thinking - a kind of estimation thinking but based on a set of evidences instead of a single research outcome; hopefully, you do some kind of meta-analysis in your review of literature
Statistical significant vs non-significant is the be-all end -all.
FALSE.
It’s a start. But it’s not the definition.
Conditional probability is the probability of an event given a certain theory.
TRUE.
Inferential statistics says, “What is the likelihood of your results in light of a certain theory?”
TRUE.
The value for the probability of that likelihood is your p-value.
P (Empirical | Theoretical) is the frequentist view of inferential statistics.
TRUE.
The sampling distribution represents the population.
TRUE.
But wait! There’s more!
We represent this distribution based on the center. BUT there is variability, naturally as come with central tendency. May standard deviation of sampling distribution. This is called “standard error”
When kurtosis is 0, that’s ideal because it’s mesokurtic.
TRUE.
We talked about the origins of standard error of the kurtosis, and standard error of the skew.
TRUE. lol
A statement that tentatively answers a research question.
Researcher’s hypotheses
Knowledge from this is best expressed in probabilities (via positivist or post-positivist)
TIP
When you do quantitative research, your researcher’s question should be answered by yes or no.
Your research question has to be “smart” or “specific enough”. The more specific, the better.
What can be deduced from the RH?
A. The Key Players: The IV and the DV (The predictor and the outcome)
B. The type of relationship between the IV and DV (is it causal or non-causal?)
C. The temporal direction of the relationship between the IV and the DV (anchor group is the gold standard; because you’re going to compare something against it; “Do boys outperform girls in algebra”? Girls is your anchor group. It’s very political, actually. The anchor group is the gold standard. There is right-tailed, left-tailed, two-tailed)
What can be deduced from the RH?
A. The Key Players: The IV and the DV (The predictor and the outcome)
B. The type of relationship between the IV and DV (is it causal or non-causal?)
C. The temporal direction of the relationship between the IV and the DV (anchor group is the gold standard; because you’re going to compare something against it; “Do boys outperform girls in algebra”? Girls is your anchor group. It’s very political, actually. The anchor group is the gold standard. There is right-tailed, left-tailed, two-tailed)
Stating the HR in a Statistical Way (In causal relationships)
- Based on the nominal levels of the IV, identify the groups you will be comparing.
- Determine which group will serve as the anchor/reference group—that is, the group that will serve as the gold standard (the ideal), such that other groups will be compared to it.
- Describe the characteristics/scope of each sample. Better if you describe:
a. sample size
b. sampling procedures
c. pertinent (controlled, similar) characteristics common between the two groups
d. the IV level on which the two groups are different - Describe the quantitative attribute of the DV in each group.
a. Measure of central tendency
b. unit of measurement (e.g. score, speed, level, frequency, percentage)
c. instrument used to measure to DV (In the example, it’s greater than, so right-tailed)
M2 (2nd element) is always your anchor group.
TRUE.
