Home
Subjects
Textbook solutions
Create
Study sets, textbooks, questions
Log in
Sign up
Upgrade to remove ads
Only $35.99/year
HSC 403- Week 8 Hypothesis Testing
STUDY
Flashcards
Learn
Write
Spell
Test
PLAY
Match
Gravity
Terms in this set (31)
What is a hypothesis test all about?
-Did our sample come from this population. . . or is that just too * improbable * ?
What is meant by saying an outcome may be too improbable?
-we know scores in the population are going to vary just by chance alone. . . some will be above average. Some will be below.
-this kind of variance = * NOISE *
Define noise
-colloquial way of describing any sort of unexplained variability
EX: if you take someone's blood pressure several times in a row, you may get slightly different values from reading to reading
We know that scores may be effected by certain variables.
These variables result in ______________________ higher scores. Or lower scores.
This kind of variance can be called _________________.
We know that scores may be effected by certain variables.
These variables result in * systematically * higher scores. Or lower scores.
This kind of variance can be called * SIGNAL * (figure on slide 13)
Give an example of signal related to blood pressure
[slide 12]
-we know certain variables (high stress, alcoholism, smoking, obesity) can lead to systematically higher blood pressures ----ppl w/ these factors should tend to have HIGHER blood pressures than the rest of the population
so, observing a higher mean than the hypothesized mean is expected
What are the two possibilities for observing an extremely high or low mean from the hypothesized mean?
1. The sample DID come from this population, and by CHANCE we just happened to get an unlikely/extreme value
2. The sample mean did NOT come from this population but from a population w/ a DIFFERENT mean
What is the main question for hypothesis testing?
"How far away from the mean is TOO far away?"
"Did our sample come from this population. . . or is that just TOO * improbable * ?"
When the data we collect differs from the population. . . what is it that a hypothesis tests will help indicate?
-a hypothesis test will say how likely this was due to chance
[Slide 13]
If a hypothesis test indicates that the likelihood of an outcome being due to chance is quite low. . . how is the explanation (outcome) interpreted?
if the chance is quite low, then we reject this explanation
A scientific method for deciding whether observed data support a theory or hypothesis is?
hypothesis testing
What type of testing uses probability to give us a quantifiable basis for testing hypotheses?
hypothesis testing
"Arjuna extract was well tolerated, but did not change LVEF (24.3 +/- 7.1 versus 25.5 +/- 7.7%; p=0.4)"
What is it you need to pay attention to in the above statement? How can you interpret?
-pay attention to p value
-we'd get difference in LVED scores of this magnitude about 40% of the time due to noise alone. . . (i.e. chance)
no good evidence to consider alternative explanations here. . . noise is too likely the reason for this difference
What are the 2 types of hypotheses
- * Null * hypothesis (H₀)
- * Alternative * hypothesis (H₁ and Ha)
Define null and alternative hypotheses
Null: your data can be explained by chance alone
Alternative: your data would unlikely be based on chance alone
and thus, there is probably a better ALTERNATIVE explanation
[slide 18, 19]
Which hypothesis do hypothesis tests examine the probability of?
-hypothesis tests examine the probability NULL hypothesis
NEVER the alternative
-"analogous to court sys. of innocent until proven guilty: the burden of evidence is on us!"
T/F: hypothesis tests examine the alternative hypothesis
false; never the alternative always the null
b/c mathematically, we're able to describe the probability of null (H₀)
Identify the H₀ and Ha hypothesis in the following:
Obj: to evaluate the changes in nursing undergraduates' levels of reflection before and after the program
H₀: the quality of undergraduates' reflections did not improve after the program
Ha: the quality of undergraduates reflections improved after the program
What is the value we use to determine how far from the mean we need to get in order to accept Ha?
the value of alpha
Define alpha
-the percent of scores for which we'll accept the Ha
-we choose alpha (aka significance level)
[slide 24]
When would an alpha level be chosen?
Describe the different types.
-researchers choose alpha level ahead of time (a priori)
-varies by field of study:
a. .01 = conservative; hard to reject H₀ the "chance" explanation
b. .05 = typical alpha level (the one we use); most often used in research
c. .10 = liberal; easier to reject H₀ the "chance" explanation . . . rarely used
How can we use alpha in SPSS to decide whether or not to reject the H₀?
-when we run a test in SPSS it will give us a column labeled "sig"
H₀ sig > alpha
Ha sig < alpha
which sig will be be greater than alpha?
the H₀ (null)
T/F: sig is the same thing as p-value
True
How would you interpret the following hypothesis test in SPSS?
"The likelihood of making an error increased with longer work hours and was three times higher when nurses worked shifts lasting of 12.5 hours or more (odds ratio = 3.29, p = .001)"
here Ha sig is LESS than alpha
(we look at p-value of .001)
slide 30
...
Reason for type 1 error α
-reason: running too many tests
our test says; Ha is true and in reality H₀ is true
Reason for type 2 error β
-reason: small sample size, not enough statistical power
our test says: H₀ is true but in reality Ha is true
-False negative
H₀: You're not preg. (to a woman)
In reality (Ha) she is pregnant
What type of error is the following:
"You're pregnant" (to a male)
-type I error; false positive
our test says; Ha is true and in reality H₀ is true
What type of error is the following?
"You're not pregnant" (to a woman who IS pregnant)
-type II error; false negative
our test says: H₀ is true but in reality Ha is true
Define statistical significance
-are the statistical results due to random chance?
Define clinical significance
-will the results matter to our patients
we will then look at effect size and generalizability (sampling approach)
Sets found in the same folder
Biostatistics for the Biological and Hea…
136 terms
HSC 403 Midterm
34 terms
HSC 403 Week 1-Introduction and Study Design
33 terms
HSC 403 Week 2-Sampling
42 terms
Other sets by this creator
Theory at a Glance
5 terms
7.4 CHES: Section VII, Area VII of Responsibility,…
22 terms
7.3 CHES: Section VII, Area VII of Responsibility,…
35 terms
7.1 CHES: Section VII, Area VII of Responsibility,…
26 terms
Other Quizlet sets
Chemistry Midterm Review
52 terms
anthro exam 2
110 terms
Principles of Econ - PREPARE Ch. 7 Questions
35 terms