According to the American Psychological Association website, validity is when a test measures what it is supposed to. Face validity is when a test outwardly appears to measure what it is intended to measure.
Criterion-related validity means that a psychological test measuring a certain trait should produce similar (correlated) results to other psychological tests measuring the trait.
Construct validity is when a test asks only about what it is supposed to measure.
Internal consistency is the property in which questions within a psychological test measure the same thing as each other and are weighted correctly within the total score.
A test should have enough questions to cover multiple aspects of a trait or disorder, without being overwhelmingly long.
A more sensitive scale will have more options for question answers and gather more specific information, while a less sensitive scale will have fewer options and gather less detailed information (e.g. one to 10 scale versus one to four scale).