HYPOTHESIS TESTING

Hypothesis means a belief about the population parameter. NOTE: Always prefer Method 3 for a one-tailed test.
 µ≠ 50,000
Denoted by Ha or H1 called as Alternate Hypothesis. CONCEPT OF ERRORS IN HYPOTHESIS TESTING
 Logical opposite or behaving figure of alternate hypothesis is
called null hypothesis denoted by H0 ieµ=50,000 H0 TRUE H0 FALSE
 If the difference of X from µ is not so significant i.e. our belief is
not correct-we reject the alternate hypothesis/ accept the null MODEL ACCEPT GOOD TYPE 2 ERROR= β=1-∝
hypothesis.
 If the difference of X from µ is significant , we need to create REJECT TYPE 1 ERROR POWER OF TEST=1-β
acceptance region around population mean. =∝
Acceptance region = µ +/- gap TYPE 1 error means rejecting a true hypothesis
= µ +/- z S.E. (Here S.E. means Standard TYPE 2 error means accepting a false hypothesis.
Error)
How to find z? Decision rules are used to minimize either Type 1 or Type 2 error.
o Level of significance-Area in tails
o One tailed test or two tailed test -Ha TESTING OF VARIANCE
o z or t Sampling distribution of variance followsChi Square distribution
Case 1: Ha µ≠ 50,000 (Two tailed test) With parameter degree of freedom =n-1
2
Case 2: Ha µ > 5,000 (One tailed test,specifically right tailed) Chi square  n-1
2 2 2
Case 3: Ha: µ<50,000(One tailed test,specifically left tailed)  statistic = (n-1)s / σ

Four methods of Hypothesis Testing for testing of Mean TESTING OF EQUALITY OF VARIANCE
 METHOD 1: ACCEPTANCE REJECTION METHOD Step 1: Define hypothesis
2 2
Step 1: Define the hypothesis Ha : σ T > σ p
2 2
Ha: µ≠ 0 H0= σ T ≤ σ P
H0: µ= 0 Step 2: One tailed test ,specifically right tailed
Step 2: Choose the significance level;nature of the test & ∝=5%
whether to use z distribution or t distribution F-distribution
2 2 2 2
Step 3: Construct the acceptance region’ Step 3: F Statistic = S 1/ S 2or S T/ S P
µ +/- t * S.E.; df= n-1 A.R. F distribution has 2 parameters i.e.
S.E. =s/√ R.R. R.R. numerator df & denominator df . For critical values, see the look
Step 4: Decision rule: Check whether X lies within the up table of F distribution; the table of F distribution gives values
acceptance [Link] the value lies outside the acceptance for probs in the right tail.
2
region,we reject the null hypothesis i.e. our belief is correct. Choose bigger variance to be S 1
Step 4: Decision Rule
 METHOD 2: CONFIDENCE INTERVAL METHOD If statistic value < critical value; we fail to reject the null.
Step 1: Define the hypothesis
Step 2: Choose the significance level;nature of the test & For a 2 tailed test;
whether to use z distribution or t distribution Step 1: Define hypothesis
2 2
Step 3: X +/- gap Ha : σ T ≠ σ p
2 2
X +/- t.S.E H0= σ T = σ P
Construct a confidence interval around X i.e. X± ∗ . . Step 2: One tailed test ,specifically right tailed
Step 4:Check whether µ lies within this interval. ∝/2=2.5%
If yes, we accept the null hypothesis, else reject it. F-distribution
2 2
Step 3: F Statistic = S T/ S P
 METHOD 3: STATISTIC CRITICAL METHOD F distribution has 2 parameters i.e.
Step 1: Define the hypothesis numerator df & denominator df .
2
Step 2: ∝(Level of significance)=5% Choose bigger variance to be S 1
Two tailed test. Try to make the test right [Link] if the test is both tailes
Step 3: z= (x-µ)/S.E. then lower critical value= 1/f at reversed degrees of freedom.
Find test statistic (t statistic) Step 4: Decision Rule
If t statistic > t critical ,(ignoring -ve signs),then reject null If F statistic lies between lower critical value & upper critical
hypothesis. value,we fail to reject the null hypothesis,else reject .

 METHOD 4: F VALUE METHOD