It turns out that Scottish New Testament scholar I. Howard Marshall (1934-2015) was a driving force behind the NIV's redefinition of adelphoi as brothers and sisters

Under I. Howard Marshall's influence, the NIV ended up interpolating "sisters" and "sister" in the New Testament about 150 times when they didn't appear in the Greek, in order to make the NT more gender neutral.

 

But to do this it had to supply a novel definition of "brothers" to justify doing so, even though the New Testament authors were perfectly capable of writing "sisters" when they wanted to, and did.

 

An example of the novel definition is shown below at Acts 1:16.

  

 

Except there is no Greek word for brothers and sisters, and it certainly isn't adelphoi.

 

To convey this meaning the language routinely relied on compounding the two words. If "brothers" were in fact the shorthand for "brothers and sisters" as the NIV insists, it is news to the gospels, which shouldn't have bothered to say "sisters" alongside "brothers" for example in Matthew 19:29, Mark 10:29f., and Luke 14:26.

 

If "brothers" means "brothers and sisters", why use "sisters" anywhere at all?

 

I. Howard Marshall (The Gospel of Luke, 1978, reprint 1986, 688) was remarkably uninstructed by the Luke 14:26 example above (If any man come to me, and hate not his father, and mother, and wife, and children, and brethren, and sisters, yea, and his own life also, he cannot be my disciple) in his interpretation of Luke 18:29 (And he said unto them, Verily I say unto you, There is no man that hath left house, or parents, or brethren, or wife, or children, for the kingdom of God's sake), where Marshall says sisters is omitted because "brothers will include both brothers and sisters just as Luke uses parents for father and mother". 

 

His authority for this is not in any early Christian literature, but in usages outside the New Testament which themselves appear only occasionally and in special contexts (e.g. when referring specifically to a brother-sister pairing, or with the classical Greek dual which virtually disappears in koine). Ordinary usage everywhere in classical and koine displays the typical male dominance of "patriarchal" society.

 

It is simply not the case in Luke 14:26 that brothers includes both brothers and sisters, nor is it justified to say brothers contains sisters in Luke 18:29 in the same way that parents contains mother and father.

 

On the contrary, Luke used parents and father and mother freely, even when parents alone would have served:

 

And when he came into the house, he suffered no man to go in, save Peter, and James, and John, and the father and the mother of the maiden. ... And her parents were astonished: but he charged them that they should tell no man what was done.

 

-- Luke 8:51, 56.

 

Marshall was straining for excuses to justify his novel definition of brothers instead of relying on New Testament and especially Luke's own usage. 

 

The fact of the matter is Luke edited out "sisters" once before when the parallels had included it (Mark 3:35 and Matthew 12:50):

 

And he answered and said unto them, My mother and my brethren are these which hear the word of God, and do it.

-- Luke 8:21.

Marshall strangely missed this opportunity at 8:21 to argue that brothers included the sisters on his new definition of adelphoi. Instead he says Luke is abbreviating Mark to avoid referring to the distasteful topic of the strife which existed between Jesus and his family. 

It's characteristic of Luke sometimes to compress language and omit elements, especially those which may seem needlessly repetitive, descriptive, or offensive to him.

Luke famously omits the scourging of Jesus by Pilate, which would have been a distasteful topic for Luke's Roman audience to read.

Luke tones down the coming judgment of Israel at the center of an eschatological drama from Matthew in favor of a gospel preached to the ends of the earth by Paul, culminating at the center of the world. "And so we came to Rome" (Acts 28:14).  

Luke omits "lands" in addition to "sisters" at 18:29, possibly as Marshall says because Luke is against the idea of material reward for disciples. Luke emphasizes that the gospel is good news for the poor, not the rich (Luke 1:53; Luke 4:18; Luke 7:22). Luke well understood that the cost of discipleship was saying goodbye to everything which one owned (Luke 14:33). Luke is nothing if not uncomfortable with the talk of rewards in Mark 10 and Matthew 19, to the point that he is completely non-committal about just what the "manifold more" actually is which the disciples will enjoy "in this present time". 

Is Luke omitting "sisters" twice in his gospel not because they are subsumed under "brothers" but rather because Luke just thinks that they are not worth mentioning? That dismissiveness could simply be unconscious expression of the "patriarchal" mindset of the times reflected more broadly in the language itself, against which feminists in the church have been raging since the 1960s.

Marshall was nothing if not a child of this time, as were the early Christians in theirs. 

Luke generally is omitting things here and there mostly because as the longest of the gospels he has to make room for a lot of other things which he wished to emphasize. Occasionally the hobby horse of a crank is a victim.

In keeping with most Greek speakers of the time, Luke does not think of females when he thinks of "brothers". He thinks of males. Marshall knew and admitted this, but Marshall's ideological prejudices overrode what he knew to be the truth: 

... Ian Howard Marshall, who was one of the translation committee members for the gender-inclusive TNIV Bible, has made an argument along these lines in a recent article:

What we are suggesting is that the usage is one in which sometimes the context may make it clear that the reference is exclusive and purely to males, but that the letters are addressed directly to mixed audiences, and therefore generally adelphoi is used in a way that does not exclude women, even if it is probable that the author may have been thinking primarily of the men. 

... But there is a problem with this argument. The problem is, the writers of the epistles ordinarily seem to be "thinking primarily of the men," as Marshall allows, and so his statement that the epistles are "addressed directly to mixed audiences" is rather problematic. If the writer is thinking primarily of the men when he addresses the congregation, then he is not really addressing the men and the women equally. Yet Marshall proceeds to build his argument concerning the meaning of adelphoi squarely on the idea that men and women are addressed equally, as equals, in the epistles of the New Testament. It is largely on the basis of this idea that a gender-neutral sense for adelphoi is posited by him and by Strauss. The whole argument collapses under the simple recognition that the apostles directed their attention primarily to the men, and used language which reflected this orientation. There is no intent to "exclude women," but the habits of speech reflect a male orientation.

... First, it is clear that Marshall's egalitarian vision of the early Church has prejudiced his whole approach to the linguistic issue at hand, in such a way that no careful philologist is likely to approve. Second, this egalitarian vision of the early Church is simply false. It represents an anachronistic illusion about the early Church, and not the historical fact of the matter. Despite the "trajectory" imagined by Marshall, men and women were not equal in the early Church. Third, Marshall ignores the contrary linguistic evidence from New Testament itself. Fourth, we note that his argument shows the inner connection of "inclusive language" with egalitarian views on the TNIV committee of translators. Spokesmen for the International Bible Society have repeatedly denied that any such ideological connection exists, and have asserted that the reasons for their gender-neutral TNIV are purely linguistic, but Marshall makes the connection clear. A revisionist egalitarian view of the early Church plays an important part in his "linguistic" argument. ...