NULL AND RESEARCH HYPOTHESES

• Scientific methodology is based on prediction.
• Hypotheses are used to predict.
• Predictions are based on “confirming” or “disconfirming” the NULL HYPOTHESIS are what we do.

NULL HYPOTHESIS is usually stated as a proposition that, if all things are random and chaotic, than no “pattern” is discernable, therefore we cannot make predictions. To the degree that patterns exist, we can make predictions and thus “fail to confirm” this assumption of randomness, chaos, or equivalence, the general lack of pattern. The null hypothesis is generally symbolized by the letter “H” followed by a subscript zero as follows::

H₀:

This failure to confirm randomness and chaos, the NULL HYPOTHIS, should result in an alternative, that there is a reliably predictable pattern. This may be described as the “Alternative Hypothesis”, sometimes called the “Research Hypothesis.” Research Hypotheses are usually based on some explanatory theory. Confirmation of an alternative or research hypothesis is suggested as a validation of the theory that generated it. The Research hypothesis is usually symbolized by the letter “H” followed by a number as follows:

H₁:

In general there are two kinds of patterns which Null Hypotheses and their alternatives examine: 1) those that predict statistically significant differences between two or more independent groups or samples; 2) those that predict statistically significant relationships between two or measures on the same sample. Each of these situations has one symbolically represented NULL HYPOTHESIS. Each of these situations may have THREE alternative RESEARCH HYPOTHESIS.

Differences

H₀:m₁ = m_{2 The two groups are the same or equivalent}

H₁: m₁ ¹ m_{2 The two groups are different but how is not predicted}

H₁: m₁ > m₂ difference between the two groups has a predictable direction! Where m₁ is greater than m₂

H₁: m₁ < m₂ difference between the two groups has a predictable direction! Where m₁ is less than m₂

Relationships

H₀: r_x.y = 0 no relationship

H₁: r_x.y ¹ 0 non-directional relationship

H₁: r_x.y > 0 Positive directional relationship (+)

H₁: r_x.y < 0 Negative directional relationship (-)