Non Probability Sampling:

Convenient sampling or accidental sampling entails the use of the most conveniently available people  as  subjects  in  the  study. Quota sampling arises from researcher's desired quota  or  proportion for  same population variable  of interest.  It  is  similar  to  stratified random sampling  in  that  the  first  step involves dividing population into homogenous strata and selecting same elements  from each of these strata. 

i)  Purposive or Judgmental Sampling involves "hand-picking"  of  subjects. The researcher might decide to purposely select widest possible variety  of respondents or might  choose  subjects who  are judged to be typical  of population  in question or knowledgeable about the issues under study. 

Purposive Sampling  is non-probability sampling  in which  the  researcher attempts to identify typical cases whatever is being studied. For example,  if you  are  studying cardiac patients, findings  "the  typical cardiac patient" would be one approach. If the  researcher is interested in  the play patterns of  four year  old village boys, then studying a small sample of "typical" four year olds attending  the "Anganwadi"  or pre-nursery schools, would comprise the purposive study sample. Obviously, this approach  is not  often a very dependable  one, for what is  "typical"  is often a very subjective opinion. Sampling  in  this subjective manner provides  no objective method for assessing  the  typicalness  of the  selected subjects. Purposive sampling should be avoided, particularly,  if  the population is' heterogeneous, and  if  they are unavoidable,  the data  analysis and interpetation  should be done carefully.