A child with an articulation disorder may have six or eight or even more different sound errors. When a speech pathologist is making a treatment plan, where does he or she begin?
First, an articulation test is given to the child. This test will assess every sound in every word position. Then the speech pathologist will determine whether the sound is stimulable – whether the child can imitate the sound. Then the sound errors will be compared to normative data to determine whether each sound should even be mastered by the child's age. For example, if a 3 year old can not pronounce / r /, that is not a concern, because 3 year olds are not expected to produce this difficult sound.
Traditionally, treatment begins with 1-4 sounds that are stimulable and early developing. When the initial sounds are mastered, treatment moves on to sounds that are later developing and / or not stimulable. There are many variables, however. For example, if a child produced a sound correctly 50% of the time, I would assume that it might be mastered unexpectedly without assistance, so I would choose a sound with which the child had more difficulty. I may choose a sound that was in the child's name or another sound important in the child's life, even though it might not fit the usual criteria.
Current research in speech pathology is leading in another direction. Some speech pathologist are advocating teaching sounds that are the most difficult, that are not stimulable, and that are not similar to each other. Although this can be more frustrating for the child and initial progress may seem slow, evidence is showing that this method results in learning transfer to sounds that are never worked with in therapy, shortening the overall time needed in therapy.
Overall, I really think that the most important point is for a child who needs speech therapy to get it. Any child who is learning and practicing new sounds will make progress, while a child who speech disorder is neglected may not improve on his own.